Arrow Research search

Author name cluster

Chi Zhang

Possible papers associated with this exact author name in Arrow. This page groups case-insensitive exact name matches and is not a full identity disambiguation profile.

81 papers
2 author rows

Possible papers

81

TMLR Journal 2026 Journal Article

ABCDE: Agentic-Based Controlled Dynamic Erasure for Intent-Aware Safety Reasoning

  • Ping Liu
  • Chi Zhang

Concept erasure has emerged as a central mechanism for safety alignment in text-conditioned generative models, yet most existing approaches implicitly adopt an unconditional suppression paradigm in which target concepts are removed whenever they appear, regardless of contextual intent. This formulation conflates benign and harmful concept usage, leading to systematic over-suppression that unnecessarily censors policy-compliant content and degrades model utility. We argue that safety intervention should instead be framed as a decision problem grounded in contextual language understanding, rather than as a purely mechanistic removal operation. Based on this perspective, we introduce Intent-Aware Concept Erasure (ICE), a decision-centric formulation that explicitly separates the question of whether a concept should be suppressed from how suppression is realized, enabling context-sensitive intervention policies that preserve benign usage while maintaining safety guarantees. To operationalize this formulation, we present Agentic-Based Controlled Dynamic Erasure (ABCDE), an agentic framework that infers a stable intervention decision from semantic context and realizes it through minimal prompt-level intervention with closed-loop multimodal output feedback, without modifying model parameters. To enable principled evaluation of intent-aware intervention, we further construct the Context-Aware Erasure Benchmark (CAEB), a paired benchmark comprising 500 prompts over 10 object concepts and 100 prompts over 5 artist styles, in which the same concept appears in both removal-required and preservation-required contexts. Experiments on CAEB show that ABCDE achieves substantially higher precision than unconditional baselines while maintaining strong recall, demonstrating effective avoidance of unnecessary suppression in benign contexts.

PDF Details

JBHI Journal 2026 Journal Article

Efficient Collaborative Model Training Mechanism With Privacy-Preserving Data for the IoMT

  • Chi Zhang
  • Tao Shen
  • Fenhua Bai
  • Xiaohui Zhang
  • Ziyuan Zhao

As time-series data from the Internet of Medical Things (IoMT) increasingly permeates various aspects of medical research, public governance, and clinical treatment, its sensitivity raises significant privacy concerns, hindering the potential of deep learning applications for cross-institutional data integration. Previous practices focused on deep learning methods based on centralized data storage and processing, which are often unsuitable for decentralized and privacy-sensitive IoMT data scenarios. Most existing methods rely on mechanisms such as trusted coordinators, which face challenges in addressing potential passive data leakage and side-channel attacks, failing to effectively protect the privacy of sensitive data during collaborative training. To address these issues, we propose a privacy-preserving collaborative training model, Secure Long Sequence Time-Series Forecasting (SecLSTF), for IoMT time-series data and design a mapping strategy between model components and Multi-Party Computation (MPC) protocols. Building on this foundation, we propose a novel secret sharing protocol, Pleione, which focuses on optimizing the computational efficiency of the low-level secret-sharing protocol. The protocol centers on a hyper-invertible matrix and adopts a paired double random expansion mechanism, significantly reducing the communication rounds required for random number generation. This optimization enhances the overall training speed of SecLSTF. Subsequently, we replace the original computational support protocol with Pleione. Experimental results show that SecLSTF-Pleione significantly reduces computational time while maintaining computational accuracy, outperforming other protocols in component efficiency. This study offers a potential pathway for cross-institutional IoMT data sharing.

Details DOI

AAAI Conference 2026 Conference Paper

Evaluating, Synthesizing, and Enhancing for Customer Support Conversation

  • Jie Zhu
  • Huaixia Dou
  • Junhui Li
  • Lifan Guo
  • Feng Chen
  • Chi Zhang
  • Fang Kong

Effective customer support requires not only accurate problem-solving but also structured and empathetic communication aligned with professional standards. However, existing dialogue datasets often lack strategic guidance, and real-world service data is difficult to access and annotate. To address this, we introduce the task of Customer Support Conversation (CSC), aimed at training customer service supporters to respond using well-defined support strategies. We propose a structured CSC framework grounded in COPC guidelines, defining five conversational stages and twelve strategies to guide high-quality interactions. Based on this, we construct CSConv, an evaluation dataset of 1,855 real-world customer–agent conversations rewritten using LLMs to reflect deliberate strategy use, and annotated accordingly. Additionally, we develop a role-playing approach that simulates strategy-rich conversations using LLM-powered roles aligned with the CSC framework, resulting in the training dataset RoleCS. Experiments show that fine-tuning strong LLMs on RoleCS significantly improves their ability to generate high-quality, strategy-aligned responses on CSConv. Human evaluations further confirm gains in problem resolution.

PDF Details DOI

AAAI Conference 2026 Conference Paper

Event-Guided Super-Resolving Blurry Image via Asymmetric Integral Driven Consistency

  • Chi Zhang
  • Xiang Zhang
  • Lei Yu
  • Gui-Song Xia
  • Yuming Fang
  • Wenhan Yang

Super-Resolution from a Blurry low-resolution image (SRB) constitutes a severely ill-posed inverse problem. Current learning-based SRB approaches primarily rely on synthetic, well-labeled paired datasets to regularize solution spaces, yet they exhibit limited generalizability in practical applications due to significant domain discrepancies between simulated degradations and real-world imaging conditions. To bridge this synthetic-to-real gap, we propose a novel Self-supervised Event-based SRB (SE-SRB) framework that leverages neuromorphic event streams as physical priors and adopts a lightweight neural architecture tailored for effective domain adaptation. Specifically, the proposed SE-SRB introduces a self-supervised learning paradigm based on asymmetric integral driven consistency, which enforces temporal coherence between predictions derived from RGB and asynchronous event streams at different time points. Extensive experiments validate that SE-SRB consistently outperforms state-of-the-art methods on both synthetic and real-world datasets. Built upon a lightweight parallel two-stream architecture, SE-SRB achieves high computational efficiency, featuring reduced parameter count, lower FLOPs, and real-time inference capability (40 FPS).

PDF Details DOI

AAAI Conference 2026 Conference Paper

Informative Subgraph Extraction with Deep Reinforcement Learning for Drug-Drug Interaction Prediction

  • Jiancong Xie
  • Wentao Wei
  • Chi Zhang
  • Jiahua Rao
  • Yuedong Yang

Drug-drug interaction (DDI) prediction is pivotal for drug safety and clinical decision-making. Recently, subgraph-based methods utilizing knowledge graphs (KGs) and domain information have achieved promising results by extracting informative subgraphs for DDI prediction. However, existing subgraph extraction methods are typically coarse-grained and nonspecific, facing two key limitations: First, they are constrained by the vast and noisy nature of real-world KGs, making it challenging to identify the most informative substructures from the massive space of candidate subgraphs. Second, current methods often fail to exploit the molecular structural specificity of drugs to selectively extract relevant subgraphs, lacking effective integration of molecular structure information with knowledge graph context. To address these challenges, we propose RISE-DDI, a novel framework for Reinforced-based Informative Subgraph Extraction approach for drug-drug interaction prediction. Specifically, RISE-DDI formulates the subgraph extraction as a Markov Decision Process (MDP) and leverages a deep reinforcement learning (RL) agent to dynamically and adaptively extract the most informative and context-specific subgraphs for each drug pair. The agent is guided by a learnable structure-aware reward model that considers both the topological context from the knowledge graph and the molecular features of the drug pairs, thereby encouraging the selection of subgraphs that are both structurally relevant and biologically informative. Extensive experiments on DDI benchmark datasets demonstrate that our method outperforms state-of-the-art baselines in both transductive and inductive scenarios, achieving improvements of up to 20%. Furthermore, visualization analyses of the extracted subgraphs highlight the interpretability of our model, providing insights into the underlying mechanisms of drug interactions.

PDF Details DOI

AAAI Conference 2026 Conference Paper

Knowledge Completes the Vision: A Multimodal Entity-aware Retrieval-Augmented Generation Framework for News Image Captioning

  • Xiaoxing You
  • Qiang Huang
  • Lingyu Li
  • Chi Zhang
  • Xiaopeng Liu
  • Min Zhang
  • Jun Yu

News image captioning aims to produce journalistically informative descriptions by combining visual content with contextual cues from associated articles. Despite recent advances, existing methods struggle with three key challenges: (1) incomplete information coverage, (2) weak cross-modal alignment, and (3) suboptimal visual-entity grounding. To address these issues, we introduce MERGE, the first Multimodal Entity-aware Retrieval-augmented GEneration framework for news image captioning. MERGE constructs an entity-centric multimodal knowledge base (EMKB) that integrates textual, visual, and structured knowledge, enabling enriched background retrieval. It improves cross-modal alignment through a multistage hypothesis-caption strategy and enhances visual-entity matching via dynamic retrieval guided by image content. Extensive experiments on GoodNews and NYTimes800k show that MERGE significantly outperforms state-of-the-art baselines, with CIDEr gains of +6.84 and +1.16 in caption quality, and F1-score improvements of +4.14 and +2.64 in named entity recognition. Notably, MERGE also generalizes well to the unseen Visual News dataset, achieving +20.17 in CIDEr and +6.22 in F1-score, demonstrating strong robustness and domain adaptability.

PDF Details DOI

AAAI Conference 2026 Conference Paper

OmniVDiff: Omni Controllable Video Diffusion for Generation and Understanding

  • Dianbing Xi
  • Jiepeng Wang
  • Yuanzhi Liang
  • Xi Qiu
  • Yuchi Huo
  • Rui Wang
  • Chi Zhang
  • Xuelong Li

In this paper, we propose a novel framework for controllable video diffusion, OmniVDiff, aiming to synthesize and comprehend multiple video visual content in a single diffusion model. To achieve this, OmniVDiff treats all video visual modalities in the color space to learn a joint distribution, while employing an adaptive control strategy that dynamically adjusts the role of each visual modality during the diffusion process, either as a generation modality or a conditioning modality. Our framework supports three key capabilities: (1) Text-conditioned video generation, where all modalities are jointly synthesized from a textual prompt; (2) Video understanding, where structural modalities are predicted from rgb inputs in a coherent manner; and (3) X-conditioned video generation, where video synthesis is guided by finegrained inputs such as depth, canny and segmentation. Extensive experiments demonstrate that OmniVDiff achieves state-of-the-art performance in video generation tasks and competitive results in video understanding. Its flexibility and scalability make it well-suited for downstream applications such as video-to-video translation, modality adaptation for visual tasks, and scene reconstruction.

PDF Details DOI

AAAI Conference 2026 Conference Paper

Streaming Generated Gaussian Process Experts for Online Learning and Control

  • Zewen Yang
  • Dongfa Zhang
  • Xiaobing Dai
  • Fengyi Yu
  • Chi Zhang
  • Bingkun Huang
  • Hamid Sadeghian
  • Sami Haddadin

Gaussian Processes (GPs), as a nonparametric learning method, offer flexible modeling capabilities and calibrated uncertainty quantification for function approximations. Additionally, GPs support online learning by efficiently incorporating new data with polynomial-time computation, making them well-suited for safety-critical dynamical systems that require rapid adaptation. However, the inference and online updates of exact GPs, when processing streaming data, incur cubic computation time and quadratic storage memory complexity, limiting their scalability to large datasets in real-time settings. In this paper, we propose a streaming kernel-induced progressively generated expert framework of Gaussian processes (SkyGP) that addresses both computational and memory constraints by maintaining a bounded set of experts, while inheriting the learning performance guarantees from exact Gaussian processes. Furthermore, two SkyGP variants are introduced, each tailored to a specific objective, either maximizing prediction accuracy (SkyGP-Dense) or improving computational efficiency (SkyGP-Fast). The effectiveness of SkyGP is validated through extensive benchmarks and real-time control experiments demonstrating its superior performance compared to state-of-the-art approaches.

PDF Details DOI

JBHI Journal 2026 Journal Article

TinnitusLLM: A Multimodal Large Language Model Framework for Tinnitus Diagnosis Through EEG-fMRI Fusion Learning

  • Yipeng Du
  • Xiaohui Chen
  • Zewei Liu
  • Zhengwu Liu
  • Ngai Wong
  • Chi Zhang
  • Jian Chen
  • Zhiwei Ding

Accurate tinnitus diagnosis is crucial for enabling timely therapeutic intervention and longitudinal treatment monitoring. While non-invasive neuroimaging modalities-particularly electroencephalography (EEG) with millisecond temporal resolution and functional magnetic resonance imaging (fMRI) with millimeter spatial resolution- provide complementary neural features, existing diagnostic approaches remain constrained to unimodal analysis of EEG or fMRI data, inherently limiting diagnostic precision and clinical generalizability. This paper introduces TinnitusLLM, the first multimodal large language model (LLM) framework that synergistically integrates EEG and fMRI features for tinnitus diagnosis. To enable LLM-based interpretation of neural signals, this framework integrates three key components: (1) a neuroinspired positional encoding mechanism that injects neurophysiological priors into the embedding space, enabling neurologically grounded, dynamic positional mapping of EEG and fMRI tokens; (2) multimodal autoregressive pretraining on more than 500 hours of EEG and 250 hours of fMRI data to learn causally informed predictive representations; and (3) fine-tuning with a cross-modal, subject-invariant adversarial learning strategy that enforces subject-independent constraints in the shared cross-modal feature space, thereby substantially improving diagnostic robustness across subjects. We validate TinnitusLLM through comprehensive experiments on a rigorously collected multimodal dataset containing 20 participants. Quantitative evaluations demonstrate that TinnitusLLM achieves superior cross-subject diagnostic accuracy compared to the state-of-the-art baseline methods. These results underscore TinnitusLLM's potential as a clinically viable framework for objective tinnitus assessment through multimodal neural decoding.

Details DOI

AAAI Conference 2026 Conference Paper

When Safe Unimodal Inputs Collide: Optimizing Reasoning Chains for Cross-Modal Safety in Multimodal Large Language Models

  • Wei Cai
  • Shujuan Liu
  • Jian Zhao
  • Ziyan Shi
  • Yusheng Zhao
  • Yuchen Yuan
  • Tianle Zhang
  • Chi Zhang

Multimodal Large Language Models (MLLMs) are susceptible to the implicit reasoning risk, wherein innocuous unimodal inputs synergistically assemble into risky multimodal data that produce harmful outputs. We attribute this vulnerability to the difficulty of MLLMs maintaining safety alignment through long-chain reasoning.To address this issue, we introduce Safe-Semantics-but-Unsafe-Interpretation (SSUI), the first dataset featuring interpretable reasoning paths tailored for such a cross-modal challenge.A novel training framework, Safety-aware Reasoning Path Optimization (SRPO), is also designed based on the SSUI dataset to align the MLLM's internal reasoning process with human safety values. Experimental results show that our SRPO-trained models achieve state-of-the-art results on key safety benchmarks, including the proposed Reasoning Path Benchmark (RSBench), significantly outperforming both open-source and top-tier commercial MLLMs.

PDF Details DOI

YNICL Journal 2026 Journal Article

White matter structure-function decoupling in juvenile myoclonic epilepsy

  • Junrui Zhang
  • Guangyao Liu
  • Yongxuan Xu
  • Chi Zhang
  • Hong Liu
  • Laiyan Ma
  • Weihao Zheng
  • Baoming Li

OBJECTIVE: Accumulating evidence highlights both structural and functional brain alterations in juvenile myoclonic epilepsy (JME), yet how these structural changes within white matter pathways drive functional disorganization remains largely unknown. Here, we aim to investigate white matter structure-function coupling (SFC) in treatment-naïve, newly diagnosed JME. METHODS: Forty-seven patients with JME and 40 demographically matched healthy controls underwent diffusion-weighted imaging (DWI) and resting-state functional magnetic resonance imaging (fMRI). Tract-wise SFC was assessed using a multivariate linear regression, with the amplitude of low-frequency fluctuations as the dependent variable and four microstructural metrics-fractional anisotropy (FA), mean diffusivity (MD), axial diffusivity (AD), and radial diffusivity (RD)-as independent variables. A support vector regression with five-fold cross-validation was employed to establish the associations with clinical severity. RESULTS: JME patients exhibited widespread white matter microstructural alterations, including increased FA and decreased diffusivity metrics, alongside functional hyperactivity in multiple tracts. Notably, a significant reduction of SFC was observed in the left corticospinal tract (P = 0.008) and left inferior longitudinal fasciculus (P = 0.006). In addition, multimodal models combining structural, functional, and coupling metrics demonstrated superior predictive performance for clinical severity compared to single-modal analyses (P = 0.026). CONCLUSION: These findings highlight white matter structure-function decoupling in the early stages of JME, specifically in key pathways relevant to motor and cognitive dysfunctions. Furthermore, the tract-specific SFC investigation offers a useful way for diagnosis, prognosis, and guiding personalized treatment strategies in this complex epilepsy syndrome.

Details DOI

NeurIPS Conference 2025 Conference Paper

Adaptive Stochastic Coefficients for Accelerating Diffusion Sampling

  • Ruoyu Wang
  • Beier Zhu
  • Junzhi Li
  • Liangyu Yuan
  • Chi Zhang

Diffusion-based generative processes, formulated as differential equation solving, frequently balance computational speed with sample quality. Our theoretical investigation of ODE- and SDE-based solvers reveals complementary weaknesses: ODE solvers accumulate irreducible gradient error along deterministic trajectories, while SDE methods suffer from amplified discretization errors when the step budget is limited. Building upon this insight, we introduce AdaSDE, a novel single-step SDE solver that aims to unify the efficiency of ODEs with the error resilience of SDEs. Specifically, we introduce a single per-step learnable coefficient, estimated via lightweight distillation, which dynamically regulates the error correction strength to accelerate diffusion sampling. Notably, our framework can be integrated with existing solvers to enhance their capabilities. Extensive experiments demonstrate state-of-the-art performance: at 5 NFE, AdaSDE achieves FID scores of $4. 18$ on CIFAR-10, $8. 05$ on FFHQ and $6. 96$ on LSUN Bedroom. Codes are available https: //github. com/WLU-wry02/AdaSDE.

PDF Details

NeurIPS Conference 2025 Conference Paper

Adversarial Locomotion and Motion Imitation for Humanoid Policy Learning

  • Jiyuan Shi
  • Xinzhe Liu
  • Dewei Wang
  • ouyang lu
  • Sören Schwertfeger
  • Chi Zhang
  • Fuchun Sun
  • Chenjia Bai

Humans exhibit diverse and expressive whole-body movements. However, attaining human-like whole-body coordination in humanoid robots remains challenging, as conventional approaches that mimic whole-body motions often neglect the distinct roles of upper and lower body. This oversight leads to computationally intensive policy learning and frequently causes robot instability and falls during real-world execution. To address these issues, we propose Adversarial Locomotion and Motion Imitation (ALMI), a novel framework that enables adversarial policy learning between upper and lower body. Specifically, the lower body aims to provide robust locomotion capabilities to follow velocity commands while the upper body tracks various motions. Conversely, the upper-body policy ensures effective motion tracking when the robot executes velocity-based movements. Through iterative updates, these policies achieve coordinated whole-body control, which can be extended to loco-manipulation tasks with teleoperation systems. Extensive experiments demonstrate that our method achieves robust locomotion and precise motion tracking in both simulation and on the full-size Unitree H1-2 robot. Additionally, we release a large-scale whole-body motion control dataset featuring high-quality episodic trajectories from MuJoCo simulations. The project page is https: //almi-humanoid. github. io.

PDF Details

NeurIPS Conference 2025 Conference Paper

BrainOmni: A Brain Foundation Model for Unified EEG and MEG Signals

  • Qinfan Xiao
  • Ziyun Cui
  • Chi Zhang
  • Siqi Chen
  • Wen Wu
  • Andrew Thwaites
  • Alexandra Woolgar
  • Bowen Zhou

Electroencephalography (EEG) and magnetoencephalography (MEG) measure neural activity non-invasively by capturing electromagnetic fields generated by dendritic currents. Although rooted in the same biophysics, EEG and MEG exhibit distinct signal patterns, further complicated by variations in sensor configurations across modalities and recording devices. Existing approaches typically rely on separate, modality- and dataset-specific models, which limits the performance and cross-domain scalability. This paper proposes BrainOmni, the first brain foundation model that generalises across heterogeneous EEG and MEG recordings. To unify diverse data sources, we introduce BrainTokenizer, the first tokeniser that quantises spatiotemporal brain activity into discrete representations. Central to BrainTokenizer is a novel Sensor Encoder that encodes sensor properties such as spatial layout, orientation, and type, enabling compatibility across devices and modalities. Building upon the discrete representations, BrainOmni learns unified semantic embeddings of brain signals by self-supervised pretraining. To the best of our knowledge, it is the first foundation model to support both EEG and MEG signals, as well as the first to incorporate large-scale MEG pretraining. A total of 1, 997 hours of EEG and 656 hours of MEG data are curated and standardised from publicly available sources for pretraining. Experiments show that BrainOmni outperforms both existing foundation models and state-of-the-art task-specific models on a range of downstream tasks. It also demonstrates strong generalisation to unseen EEG and MEG devices. Further analysis reveals that joint EEG-MEG (EMEG) training yields consistent improvements across both modalities. Code and checkpoints are publicly available at https: //github. com/OpenTSLab/BrainOmni

PDF Details

NeurIPS Conference 2025 Conference Paper

CAS-Spec: Cascade Adaptive Self-Speculative Decoding for On-the-Fly Lossless Inference Acceleration of LLMs

  • Zhiyuan Ning
  • Jiawei Shao
  • Ruge Xu
  • Xinfei Guo
  • Jun Zhang
  • Chi Zhang
  • Xuelong Li

Speculative decoding has become a widely adopted as an effective technique for lossless inference acceleration when deploying large language models (LLMs). While on-the-fly self-speculative methods offer seamless integration and broad utility, they often fall short of the speed gains achieved by methods relying on specialized training. Cascading a hierarchy of draft models promises further acceleration and flexibility, but the high cost of training multiple models has limited its practical application. In this paper, we propose a novel Cascade Adaptive Self-Speculative Decoding (CAS-Spec) method which constructs speculative draft models by leveraging dynamically switchable inference acceleration (DSIA) strategies, including layer sparsity and activation quantization. Furthermore, traditional vertical and horizontal cascade algorithms are inefficient when applied to self-speculative decoding methods. We introduce a Dynamic Tree Cascade (DyTC) algorithm that adaptively routes the multi-level draft models and assigns the draft lengths, based on the heuristics of acceptance rates and latency prediction. Our CAS-Spec method achieves state-of-the-art acceleration compared to existing on-the-fly speculative decoding methods, with an average speedup from $1. 1\times$ to $2. 3\times$ over autoregressive decoding across various LLMs and datasets. DyTC improves the average speedup by $47$\% and $48$\% over cascade-based baseline and tree-based baseline algorithms, respectively. CAS-Spec can be easily integrated into most existing LLMs and holds promising potential for further acceleration as self-speculative decoding techniques continue to evolve.

PDF Details

UAI Conference 2025 Conference Paper

Creative Agents: Empowering Agents with Imagination for Creative Tasks

  • Penglin Cai
  • Chi Zhang
  • Yuhui Fu 0004
  • Haoqi Yuan
  • Zongqing Lu 0002

We study building embodied agents for open-ended creative tasks. While existing methods build instruction-following agents that can perform diverse open-ended tasks, none of them demonstrates creativity – the ability to give novel and diverse solutions implicit in the language instructions. This limitation comes from their inability to convert abstract language instructions into concrete goals and perform long-horizon planning for such complicated goals. Given the observation that humans perform creative tasks with imagination, we propose a class of solutions, where the controller is enhanced with an imaginator generating detailed imaginations of task outcomes conditioned on language instructions. We introduce several approaches to implementing the components of creative agents. We implement the imaginator with either a large language model for textual imagination or a diffusion model for visual imagination. The controller can either be a behavior-cloning policy or a pre-trained foundation model generating executable codes in the environment. We benchmark creative tasks with the challenging open-world game Minecraft, where the agents create diverse buildings given free-form language instructions. We propose novel evaluation metrics for open-ended creative tasks utilizing GPT-4V, which holds many advantages over existing metrics. We perform a detailed experimental analysis of creative agents, showing that creative agents are the first AI agents accomplishing diverse building creation in the survival mode of Minecraft. Our benchmark and models are open-source for future research on creative agents (https: //github. com/PKU-RL/Creative-Agents).

Details

NeurIPS Conference 2025 Conference Paper

DAPO: An Open-Source LLM Reinforcement Learning System at Scale

  • Qiying Yu
  • Zheng Zhang
  • Ruofei Zhu
  • Yufeng Yuan
  • Xiaochen Zuo
  • Yu Yue
  • Weinan Dai
  • Tiantian Fan

Inference scaling empowers LLMs with unprecedented reasoning ability, with reinforcement learning as the core technique to elicit complex reasoning. However, key technical details of state-of-the-art reasoning LLMs are concealed (such as in OpenAI o1 blog and DeepSeek R1 technical report), thus the community still struggles to reproduce their RL training results. We propose the D ecoupled Clip and D ynamic s A mpling P olicy O ptimization ( DAPO ) algorithm, and fully open-source a state-of-the-art large-scale RL system that achieves 50 points on AIME 2024 using Qwen2. 5-32B base model. Unlike previous works that withhold training details, we introduce four key techniques of our algorithm that make large-scale LLM RL a success. In addition, we open-source our training code, which is built on the verl framework, along with a carefully curated and processed dataset. These components of our open-source system enhance reproducibility and support future research in large-scale LLM RL.

PDF Details

AAAI Conference 2025 Conference Paper

DualOpt: A Dual Divide-and-Optimize Algorithm for the Large-scale Traveling Salesman Problem

  • Shipei Zhou
  • Yuandong Ding
  • Chi Zhang
  • Zhiguang Cao
  • Yan Jin

This paper proposes a dual divide-and-optimize algorithm (DualOpt) for solving the large-scale traveling salesman problem (TSP). DualOpt combines two complementary strategies to improve both solution quality and computational efficiency. The first strategy is a grid-based divide-and-conquer procedure that partitions the TSP into smaller sub-problems, solving them in parallel and iteratively refining the solution by merging nodes and partial routes. The process continues until only one grid remains, yielding a high-quality initial solution. The second strategy involves a path-based divide-and-optimize procedure that further optimizes the solution by dividing it into sub-paths, optimizing each using a neural solver, and merging them back to progressively improve the overall solution. Extensive experiments conducted on two groups of TSP benchmark instances, including randomly generated instances with up to 100,000 nodes and real-world datasets from TSPLIB, demonstrate the effectiveness of DualOpt. The proposed DualOpt achieves highly competitive results compared to 10 state-of-the-art algorithms in the literature. In particular, DualOpt achieves an improvement gap up to 1.40% for the largest instance TSP100K with a remarkable 104x speed-up over the leading heuristic solver LKH3. Additionally, DualOpt demonstrates strong generalization on TSPLIB benchmarks, confirming its capability to tackle diverse real-world TSP applications.

PDF Details DOI

ICML Conference 2025 Conference Paper

From Weight-Based to State-Based Fine-Tuning: Further Memory Reduction on LoRA with Parallel Control

  • Chi Zhang
  • Lianhai Ren
  • Jingpu Cheng
  • Qianxiao Li

The LoRA method has achieved notable success in reducing GPU memory usage by applying low-rank updates to weight matrices. Yet, one simple question remains: can we push this reduction even further? Furthermore, is it possible to achieve this while improving performance and reducing computation time? Answering these questions requires moving beyond the conventional weight-centric approach. In this paper, we present a state-based fine-tuning framework that shifts the focus from weight adaptation to optimizing forward states, with LoRA acting as a special example. Specifically, state-based tuning introduces parameterized perturbations to the states within the computational graph, allowing us to control states across an entire residual block. A key advantage of this approach is the potential to avoid storing large intermediate states in models like transformers. Empirical results across multiple architectures—including ViT, RoBERTa, LLaMA2-7B, and LLaMA3-8B—show that our method further reduces memory consumption and computation time while simultaneously improving performance. Moreover, as a result of memory reduction, we explore the feasibility to train 7B/8B models on consumer-level GPUs like Nvidia 3090, without model quantization. The code is available at an anonymous GitHub repository

Details

NeurIPS Conference 2025 Conference Paper

Generalized and Invariant Single-Neuron In-Vivo Activity Representation Learning

  • Wei Wu
  • Yuxing Lu
  • Zhengrui Guo
  • Chi Zhang
  • Can Liao
  • Yifan Bu
  • Fangxu Zhou
  • Jinzhuo Wang

In computational neuroscience, models representing single-neuron in-vivo activity have become essential for understanding the functional identities of individual neurons. These models, such as implicit representation methods based on Transformer architectures, contrastive learning frameworks, and variational autoencoders, aim to capture the invariant and intrinsic computational features of single neurons. The learned single-neuron computational role representations should remain invariant across changing environment and are affected by their molecular expression and location. Thus, the representations allow for in vivo prediction of the molecular cell types and anatomical locations of single neurons, facilitating advanced closed-loop experimental designs. However, current models face the problem of limited generalizability. This is due to batch effects caused by differences in experimental design, animal subjects, and recording platforms. These confounding factors often lead to overfitting, reducing the robustness and practical utility of the models across various experimental scenarios. Previous studies have not rigorously evaluated how well the models generalize to new animals or stimulus conditions, creating a significant gap in the field. To solve this issue, we present a comprehensive experimental protocol that explicitly evaluates model performance on unseen animals and stimulus types. Additionally, we propose a model-agnostic adversarial training strategy. In this strategy, a discriminator network is used to eliminate batch-related information from the learned representations. The adversarial framework forces the representation model to focus on the intrinsic properties of neurons, thereby enhancing generalizability. Our approach is compatible with all major single-neuron representation models and significantly improves model robustness. This work emphasizes the importance of generalization in single-neuron representation models and offers an effective solution, paving the way for the practical application of computational models in vivo. It also shows potential for building unified atlases based on single-neuron in vivo activity.

PDF Details

NeurIPS Conference 2025 Conference Paper

Handling Label Noise via Instance-Level Difficulty Modeling and Dynamic Optimization

  • Kuan Zhang
  • Chengliang Chai
  • Jingzhe Xu
  • Chi Zhang
  • Han Han
  • Ye Yuan
  • Guoren Wang
  • Lei Cao

Recent studies indicate that deep neural networks degrade in generalization performance under noisy supervision. Existing methods focus on isolating clean subsets or correcting noisy labels, facing limitations such as high computational costs, heavy hyperparameter tuning process, and coarse-grained optimization. To address these challenges, we propose a novel two-stage noisy learning framework that enables instance-level optimization through a dynamically weighted loss function, avoiding hyperparameter tuning. To obtain stable and accurate information about noise modeling, we introduce a simple yet effective metric, termed $\textit{wrong event}$, which dynamically models the cleanliness and difficulty of individual samples while maintaining computational costs. Our framework first collects $\textit{wrong event}$ information and builds a strong base model. Then we perform noise-robust training on the base model, using a probabilistic model to handle the $\textit{wrong event}$ information of samples. Experiments on six synthetic and real-world LNL benchmarks demonstrate our method surpasses state-of-the-art methods in performance, achieves a nearly 75\% reduction in storage and computational time, strongly improving model scalability. Our code is available at https: //github. com/iTheresaApocalypse/IDO.

PDF Details

NeurIPS Conference 2025 Conference Paper

Heterogeneous Adversarial Play in Interactive Environments

  • Manjie Xu
  • Xinyi Yang
  • Jiayu Zhan
  • Wei Liang
  • Chi Zhang
  • Yixin Zhu

Self-play constitutes a fundamental paradigm for autonomous skill acquisition, whereby agents iteratively enhance their capabilities through self-directed environmental exploration. Conventional self-play frameworks exploit agent symmetry within zero-sum competitive settings, yet this approach proves inadequate for open-ended learning scenarios characterized by inherent asymmetry. Human pedagogical systems exemplify asymmetric instructional frameworks wherein educators systematically construct challenges calibrated to individual learners' developmental trajectories. The principal challenge resides in operationalizing these asymmetric, adaptive pedagogical mechanisms within artificial systems capable of autonomously synthesizing appropriate curricula without predetermined task hierarchies. Here we present Heterogeneous Adversarial Play (HAP), an adversarial Automatic Curriculum Learning framework that formalizes teacher-student interactions as a minimax optimization wherein task-generating instructor and problem-solving learner co-evolve through adversarial dynamics. In contrast to prevailing automatic curriculum learning methodologies that employ static curricula or unidirectional task selection mechanisms, HAP establishes a bidirectional feedback system wherein instructors continuously recalibrate task complexity in response to real-time learner performance metrics. Experimental validation across multi-task learning domains demonstrates that our framework achieves performance parity with SOTA baselines while generating curricula that enhance learning efficacy in both artificial agents and human subjects.

PDF Details

AAAI Conference 2025 Conference Paper

Image Regeneration: Evaluating Text-to-Image Model via Generating Identical Image with Multimodal Large Language Models

  • Chutian Meng
  • Fan Ma
  • Jiaxu Miao
  • Chi Zhang
  • Yi Yang
  • Yueting Zhuang

Diffusion models have revitalized the image generation domain, playing crucial roles in both academic research and artistic expression. With the emergence of new diffusion models, assessing the performance of text-to-image models has become increasingly important. Current metrics focus on directly matching the input text with the generated image, but due to cross-modal information asymmetry, this leads to unreliable or incomplete assessment results. Motivated by this, we introduce the Image Regeneration task in this study to assess text-to-image models by tasking the T2I model with generating an image according to the reference image. We use GPT4V to bridge the gap between the reference image and the text input for the T2I model, allowing T2I models to understand image content. This evaluation process is simplified as comparisons between the generated image and the reference image are straightforward. Two regeneration datasets spanning content-diverse and style-diverse evaluation dataset are introduced to evaluate the leading diffusion models currently available. Additionally, we present ImageRepainter framework to enhance the quality of generated images by improving content comprehension via MLLM guided iterative generation and revision. Our comprehensive experiments have showcased the effectiveness of this framework in assessing the generative capabilities of models. By leveraging MLLM, we have demonstrated that a robust T2M can produce images more closely resembling the reference image.

PDF Details DOI

NeurIPS Conference 2025 Conference Paper

KINDLE: Knowledge-Guided Distillation for Prior-Free Gene Regulatory Network Inference

  • Rui Peng
  • Yuchen Lu
  • Qichen Sun
  • Yuxing Lu
  • Chi Zhang
  • Ziru Liu
  • Jinzhuo Wang

Gene regulatory network (GRN) inference serves as a cornerstone for deciphering cellular decision-making processes. Early approaches rely exclusively on gene expression data, thus their predictive power remain fundamentally constrained by the vast combinatorial space of potential gene-gene interactions. Subsequent methods integrate prior knowledge to mitigate this challenge by restricting the solution space to biologically plausible interactions. However, we argue that the effectiveness of these approaches is contingent upon the precision of prior information and the reduction in the search space will circumscribe the models' potential for novel biological discoveries. To address these limitations, we introduce KINDLE, a three-stage framework that decouples GRN inference from prior knowledge dependencies. KINDLE trains a teacher model that integrates prior knowledge with temporal gene expression dynamics and subsequently distills this encoded knowledge to a student model, enabling accurate GRN inference solely from expression data without access to any prior. KINDLE achieves state-of-the-art performance across four benchmark datasets. Notably, it successfully identifies key transcription factors governing mouse embryonic development and precisely characterizes their functional roles. In mouse hematopoietic stem cell data, KINDLE accurately predicts fate transition outcomes following knockout of two critical regulators (Gata1 and Spi1). These biological validations demonstrate our framework's dual capability in maintaining topological inference precision while preserving discovery potential for novel biological mechanisms.

PDF Details

JBHI Journal 2025 Journal Article

MedTsLLM: Medical Time Series Analysis Using Multimodal LLMs

  • Nimeesha Chan
  • Felix Parker
  • Chi Zhang
  • William Bennett
  • Mung Yao Jia
  • James Fackler
  • Kimia Ghobadi

Traditional machine learning approaches for biomedical time series analysis face fundamental limitations when integrating the heterogeneous data types essential for comprehensive clinical understanding. Physiological signals must be interpreted within rich clinical contexts that include patient history, current medications, and treatment protocols-information typically stored as unstructured text that conventional time series models cannot effectively utilize. We propose MedTsLLM, a multimodal model that aims to address this critical gap by integrating numerical physiological signals with natural language clinical information through large language models (LLMs). Our framework incorporates patch reprogramming for time series-LLM alignment and introduces two key innovations: novel covariate handling strategies that capture complex physiological relationships, and contextual prompting mechanisms that incorporate patient-specific information. MedTsLLM addresses four clinically significant tasks within a unified architecture: semantic segmentation, boundary detection, anomaly detection, and classification. Through comprehensive evaluation across diverse medical domains, including ECG analysis, respiratory monitoring, and cardiac arrhythmia detection, our approach consistently outperforms state-of-the-art baselines across all tasks and datasets. These results demonstrate the transformative potential of multimodal LLMs for biomedical signal analysis, enabling clinicians to extract deeper insights from physiological data while leveraging comprehensive clinical context to enhance diagnostic accuracy, patient monitoring, and personalized treatment decisions.

Details DOI

ICLR Conference 2025 Conference Paper

Model-Free Offline Reinforcement Learning with Enhanced Robustness

  • Chi Zhang
  • Zain Ulabedeen Farhat
  • George K. Atia
  • Yue Wang 0068

Offline reinforcement learning (RL) has gained considerable attention for its ability to learn policies from pre-collected data without real-time interaction, which makes it particularly useful for high-risk applications. However, due to its reliance on offline datasets, existing works inevitably introduce assumptions to ensure effective learning, which, however, often lead to a trade-off between robustness to model mismatch and scalability to large environments. In this paper, we enhance both aspects with a novel double-pessimism principle, which conservatively estimates performance and accounts for both limited data and potential model mismatches, two major reasons for the previous trade-off. We then propose a universal, model-free algorithm to learn a policy that is robust to potential environment mismatches, which enhances robustness in a scalable manner. Furthermore, we provide a sample complexity analysis of our algorithm when the mismatch is modeled by the $l_\alpha$-norm, which also theoretically demonstrates the efficiency of our method. Extensive experiments further demonstrate that our approach significantly improves robustness in a more scalable manner than existing methods.

Details

NeurIPS Conference 2025 Conference Paper

NFIG: Multi-Scale Autoregressive Image Generation via Frequency Ordering

  • Zhihao Huang
  • Xi Qiu
  • Yukuo Ma
  • Yifu Zhou
  • Junjie Chen
  • Hongyuan Zhang
  • Chi Zhang
  • Xuelong Li

Autoregressive models have achieved significant success in image generation. However, unlike the inherent hierarchical structure of image information in the spectral domain, standard autoregressive methods typically generate pixels sequentially in a fixed spatial order. To better leverage this spectral hierarchy, we introduce Next-Frequency Image Generation (NFIG). NFIG is a novel framework that decomposes the image generation process into multiple frequency-guided stages. NFIG aligns the generation process with the natural image structure. It does this by first generating low-frequency components, which efficiently capture global structure with significantly fewer tokens, and then progressively adding higher-frequency details. This frequency-aware paradigm offers substantial advantages: it not only improves the quality of generated images but crucially reduces inference cost by efficiently establishing global structure early on. Extensive experiments on the ImageNet-256 benchmark validate NFIG's effectiveness, demonstrating superior performance (FID: 2. 81) and a notable 1. 25x speedup compared to the strong baseline VAR-d20.

PDF Details

ICML Conference 2025 Conference Paper

Pessimism Principle Can Be Effective: Towards a Framework for Zero-Shot Transfer Reinforcement Learning

  • Chi Zhang
  • Ziying Jia
  • George K. Atia
  • Sihong He
  • Yue Wang 0068

Transfer reinforcement learning aims to derive a near-optimal policy for a target environment with limited data by leveraging abundant data from related source domains. However, it faces two key challenges: the lack of performance guarantees for the transferred policy, which can lead to undesired actions, and the risk of negative transfer when multiple source domains are involved. We propose a novel framework based on the pessimism principle, which constructs and optimizes a conservative estimation of the target domain’s performance. Our framework effectively addresses the two challenges by providing an optimized lower bound on target performance, ensuring safe and reliable decisions, and by exhibiting monotonic improvement with respect to the quality of the source domains, thereby avoiding negative transfer. We construct two types of conservative estimations, rigorously characterize their effectiveness, and develop efficient distributed algorithms with convergence guarantees. Our framework provides a theoretically sound and practically robust solution for transfer learning in reinforcement learning.

Details

NeurIPS Conference 2025 Conference Paper

RoboRefer: Towards Spatial Referring with Reasoning in Vision-Language Models for Robotics

  • Enshen Zhou
  • Jingkun An
  • Cheng Chi
  • Yi Han
  • Shanyu Rong
  • Chi Zhang
  • Pengwei Wang
  • Zhongyuan Wang

Spatial referring is a fundamental capability of embodied robots to interact with the 3D physical world. However, even with the powerful pretrained VLMs, recent approaches are still not qualified to accurately understand the complex 3D scenes and dynamically reason about the instruction-indicated locations for interaction. To this end, we propose RoboRefer, a 3D-aware vision language model (VLM) that can first achieve precise spatial understanding by integrating a disentangled but dedicated depth encoder via supervised fine-tuning (SFT). Moreover, RoboRefer advances generalized multi-step spatial reasoning via reinforcement fine-tuning (RFT), with metric-sensitive process reward functions tailored for spatial referring tasks. To support SFT and RFT training, we introduce RefSpatial, a large-scale dataset of 20M QA pairs (2x prior), covering 31 spatial relations (vs. 15 prior) and supporting complex reasoning processes (up to 5 steps). In addition, we introduce RefSpatial-Bench, a challenging benchmark filling the gap in evaluating spatial referring with multi-step reasoning. Experiments show that SFT-trained RoboRefer achieves state-of-the-art spatial understanding, with an average success rate of 89. 6%. RFT-trained RoboRefer further outperforms all other baselines by a large margin, even surpassing Gemini-2. 5-Pro by 12. 4% in average accuracy on RefSpatial-Bench. Notably, RoboRefer can be integrated with various control policies to execute long-horizon, dynamic tasks across diverse robots (e, g. , UR5, G1 humanoid) in cluttered real-world scenes.

PDF Details

AAAI Conference 2025 Conference Paper

Sim4Rec: Data-Free Model Extraction Attack on Sequential Recommendation

  • Yihao Wang
  • Jiajie Su
  • Chaochao Chen
  • Meng Han
  • Chi Zhang
  • Jun Wang

Model extraction attack shows promising performance in revealing sequential recommendation (SeqRec) robustness, e.g., as an upstream task of transfer-based attack to provide optimization feedback for downstream attacks. However, existing work either heavily relies on impractical prior knowledge or has impressive attack performance. In this paper, we focus on data-free model extraction attack on SeqRec, which aims to efficiently train a surrogate model that closely imitates the target model in a practical setting. Conducting such an attack is challenging. First, imitating sequential training data for accurate model extraction is hard without prior knowledge. Second, limited queries for the target model require the attack to be efficient. To address these challenges, we propose a novel adversarial framework Sim4Rec which includes two modules, i.e., controllable sequence generation and reinforced adversarial distillation. The former allows a sequential generator to produce synthetic data similar to training data through pre-training with controllable generated samples. The latter efficiently extracts the target model via reinforced adversarial knowledge distillation. Extensive experiments demonstrate the advancement of Sim4Rec.

PDF Details DOI

AAAI Conference 2025 Conference Paper

Style Nursing with Spatial and Semantic Guidance for Zero-Shot Traffic Scene Style Transfer

  • Zhen Wang
  • Zihang Lin
  • Meng Yuan
  • Yuehu Liu
  • Chi Zhang

Recent advances in text-to-image diffusion models have shown an outstanding ability in zero-shot style transfer. However, existing methods often struggle to balance preserving the semantic content of the input image and faithfully transferring the target style in line with the edit prompt. Especially when applied to complex traffic scenes with diverse objects, layouts, and stylistic variations, current diffusion models tend to exhibit Style Neglection, i.e., failing to generate the required style in the prompt. To address this issue, we propose Style Nursing, which directs the model to focus on style subject tokens in the text prompt and excites their corresponding visual activations. Moreover, we introduce Spatial and Semantic Guidance to guide the preservation of content after editing, which utilizes spatial features from the DDIM sampling process together with attention maps from the semantic reconstruction. To evaluate the performance of zero-shot style transfer methods in traffic scenes, we present STREET-6K, a new benchmark dataset comprising 6000 images showcasing diverse traffic scenes and style transfer variations, accompanied by comprehensive annotations and evaluation metrics. Our approach beats state-of-the-art image translation methods in comprehensive quantitative metrics and human evaluations on traffic scene image synthesis while seamlessly generalizing to various other types of images without training or fine-tuning. Further experiments on detection and segmentation tasks show that fine-tuning perception models on our synthesized images improves Recall and mean Intersection over Union (mIoU) by over 10% and 3% respectively in rarely-seen traffic scenes.

PDF Details DOI

AAAI Conference 2025 Conference Paper

SubjectDrive: Scaling Generative Data in Autonomous Driving via Subject Control

  • Binyuan Huang
  • Yuqing Wen
  • Yucheng Zhao
  • Yaosi Hu
  • Yingfei Liu
  • Fan Jia
  • Weixin Mao
  • Tiancai Wang

Autonomous driving progress relies on large-scale annotated datasets. In this work, we explore the potential of generative models to produce vast quantities of freely-labeled data for autonomous driving applications and present SubjectDrive, the first model proven to scale generative data production in a way that could continuously improve autonomous driving applications. We investigate the impact of scaling up the quantity of generative data on the performance of downstream perception models and find that enhancing data diversity plays a crucial role in effectively scaling generative data production. Therefore, we have developed a novel model equipped with a subject control mechanism, which allows the generative model to leverage diverse external data sources for producing varied and useful data. Extensive evaluations confirm SubjectDrive's efficacy in generating scalable autonomous driving training data, marking a significant step toward revolutionizing data production methods in this field.

PDF Details DOI

NeurIPS Conference 2025 Conference Paper

SynLogic: Synthesizing Verifiable Reasoning Data at Scale for Learning Logical Reasoning and Beyond

  • Junteng Liu
  • Yuanxiang Fan
  • Jiang Zhuo
  • Han Ding
  • Yongyi Hu
  • Chi Zhang
  • Yiqi Shi
  • Shitong Weng

Recent advances such as OpenAI-o1 and DeepSeek R1 have demonstrated the potential of Reinforcement Learning (RL) to enhance reasoning abilities in Large Language Models (LLMs). While open-source replication efforts have primarily focused on mathematical and coding domains, methods and resources for developing general reasoning capabilities remain underexplored. This gap is partly due to the challenge of collecting diverse and verifiable reasoning data suitable for RL. We hypothesize that logical reasoning is critical for developing general reasoning capabilities, as logic forms a fundamental building block of reasoning. In this work, we present SynLogic, a data synthesis framework and dataset that generates diverse logical reasoning data at scale, encompassing 35 diverse logical reasoning tasks. The SynLogic approach enables controlled synthesis of data with adjustable difficulty and quantity. Importantly, all examples can be verified by simple rules, making them ideally suited for RL with verifiable rewards. In our experiments, we validate the effectiveness of RL training on the SynLogic dataset based on 7B and 32B models. SynLogic leads to state-of-the-art logical reasoning performance among open-source datasets, surpassing DeepSeek-R1-Distill-Qwen-32B by 6 points on BBEH. Furthermore, mixing SynLogic data with mathematical and coding tasks improves the training efficiency of these domains and significantly enhances reasoning generalization. Notably, our mixed training model outperforms DeepSeek-R1-Zero-Qwen-32B across multiple benchmarks. These findings position SynLogic as a valuable resource for advancing the broader reasoning capabilities of LLMs. We will open-source both the data synthesis pipeline and the SynLogic dataset.

PDF Details

NeurIPS Conference 2025 Conference Paper

Towards Reliable LLM-based Robots Planning via Combined Uncertainty Estimation

  • Shiyuan Yin
  • Chenjia Bai
  • Zihao Zhang
  • Junwei Jin
  • Xinxin Zhang
  • Chi Zhang
  • Xuelong Li

Large language models (LLMs) demonstrate advanced reasoning abilities, enabling robots to understand natural language instructions and generate high-level plans with appropriate grounding. However, LLM hallucinations present a significant challenge, often leading to overconfident yet potentially misaligned or unsafe plans. While researchers have explored uncertainty estimation to improve the reliability of LLM-based planning, existing studies have not sufficiently differentiated between epistemic and intrinsic uncertainty, limiting the effectiveness of uncertainty estimation. In this paper, we present Combined Uncertainty estimation for Reliable Embodied planning (CURE), which decomposes the uncertainty into epistemic and intrinsic uncertainty, each estimated separately. Furthermore, epistemic uncertainty is subdivided into task clarity and task familiarity for more accurate evaluation. The overall uncertainty assessments are obtained using random network distillation and multi-layer perceptron regression heads driven by LLM features. We validated our approach in two distinct experimental settings: kitchen manipulation and tabletop rearrangement experiments. The results show that, compared to existing methods, our approach yields uncertainty estimates that are more closely aligned with the actual execution outcomes. The code is at https: //github. com/Firesuiry/CURE.

PDF Details

JBHI Journal 2025 Journal Article

Valence-Arousal Disentangled Representation Learning for Emotion Recognition in SSVEP-Based BCIs

  • Yipeng Du
  • Jie Chen
  • Zhengwu Liu
  • Ngai Wong
  • Chi Zhang
  • Zhiwei Ding
  • Jian Liu
  • Edith C.H. Ngai

Steady state visually evoked potential (SSVEP)-based brain-computer interfaces (BCIs), which are widely used in rehabilitation and disability assistance, can benefit from real-time emotion recognition to enhance human–machine interaction. However, the learned discri-minative latent representations in SSVEP-BCIs may generalize in an unintended direction, which can lead to reduced accuracy in detecting emotional states. In this paper, we introduce a Valence-Arousal Disentangled Representation Learning (VADL) method, drawing inspir-ation from the classical two-dimensional emotional model, to enhance the performance and generalization of emotion recognition within SSVEP-BCIs. VADL distinctly disentangles the latent variables of valence and arousal information to improve accuracy. It utilizes the structured state space duality model to thoroughly extract global emotional features. Additionally, we propose a Multisubject Gradient Blending training strategy that individually tailors the learning pace of reconstruction and discrimination tasks within VADL on-the-fly. To verify the feasibility of our method, we have developed a comprehensive database comprising 23 subjects, in which both the emotional states and SSVEPs were effectively elicited. Experimental results indicate that VADL surpasses existing state-of-the-art benchmark algorithms.

Details DOI

NeurIPS Conference 2025 Conference Paper

VLM in a flash: I/O-Efficient Sparsification of Vision-Language Model via Neuron Chunking

  • Kichang Yang
  • Seonjun Kim
  • Minjae Kim
  • Nairan Zhang
  • Chi Zhang
  • Youngki Lee

Edge deployment of large Vision-Language Models (VLMs) increasingly relies on flash-based weight offloading, where activation sparsification is used to reduce I/O overhead. However, conventional sparsification remains model-centric, selecting neurons solely by activation magnitude and neglecting how access patterns influence flash performance. We present Neuron Chunking, an I/O-efficient sparsification strategy that operates on chunks —groups of contiguous neurons in memory—and couples neuron importance with storage access cost. The method models I/O latency through a lightweight abstraction of access contiguity and selects chunks with high utility, defined as neuron importance normalized by estimated latency. By aligning sparsification decisions with the underlying storage behavior, Neuron Chunking improves I/O efficiency by up to 4. 65× and 5. 76× on Jetson Orin Nano and Jetson AGX Orin, respectively. The code is available at https: //github. com/snuhcs/vlm-flash.

PDF Details

AAAI Conference 2024 Conference Paper

Adaptive Hardness Negative Sampling for Collaborative Filtering

  • Riwei Lai
  • Rui Chen
  • Qilong Han
  • Chi Zhang
  • Li Chen

Negative sampling is essential for implicit collaborative filtering to provide proper negative training signals so as to achieve desirable performance. We experimentally unveil a common limitation of all existing negative sampling methods that they can only select negative samples of a fixed hardness level, leading to the false positive problem (FPP) and false negative problem (FNP). We then propose a new paradigm called adaptive hardness negative sampling (AHNS) and discuss its three key criteria. By adaptively selecting negative samples with appropriate hardnesses during the training process, AHNS can well mitigate the impacts of FPP and FNP. Next, we present a concrete instantiation of AHNS called AHNS_{p<0}, and theoretically demonstrate that AHNS_{p<0} can fit the three criteria of AHNS well and achieve a larger lower bound of normalized discounted cumulative gain. Besides, we note that existing negative sampling methods can be regarded as more relaxed cases of AHNS. Finally, we conduct comprehensive experiments, and the results show that AHNS_{p<0} can consistently and substantially outperform several state-of-the-art competitors on multiple datasets.

PDF Details DOI

EAAI Journal 2024 Journal Article

An active semi-supervised transfer learning method for robot pose error prediction and compensation

  • Teng Zhang
  • Fangyu Peng
  • Xiaowei Tang
  • Rong Yan
  • Chi Zhang
  • Runpeng Deng

Robots are widely employed in industrial settings owing to their efficiency, flexibility, and extensive operational ranges. However, their application in high-precision scenarios is limited owing to their low absolute accuracies. Existing methods suffer from high measurement costs, and limited applicability and accuracy. To address these issues, an active semi supervised transfer learning method (ASTL) is introduced. The pose error prediction problem was modelled as a transfer learning paradigm for the first time. It leverages the proposed multi-stage greedy sampling (MGS) for informed sample labelling combined with coarse calibration and semi supervised transfer learning (STL) to embed theoretical knowledge for globally accurate predictions. The proposed method is compared with other prediction and compensation approaches for four robotic motion tasks. It significantly reduces the time consumption by approximately 89. 3% compared with direct measurements and achieves a maximum reduction of approximately 90% in robot pose errors. This substantial enhancement in pose accuracy promotes high-quality applications of robots in the industrial sector.

Details DOI

NeurIPS Conference 2024 Conference Paper

Do Finetti: On Causal Effects for Exchangeable Data

  • Siyuan Guo
  • Chi Zhang
  • Karthika Mohan
  • Ferenc Huszár
  • Bernhard Schölkopf

We study causal effect estimation in a setting where the data are not i. i. d. $\ $(independent and identically distributed). We focus on exchangeable data satisfying an assumption of independent causal mechanisms. Traditional causal effect estimation frameworks, e. g. , relying on structural causal models and do-calculus, are typically limited to i. i. d. data and do not extend to more general exchangeable generative processes, which naturally arise in multi-environment data. To address this gap, we develop a generalized framework for exchangeable data and introduce a truncated factorization formula that facilitates both the identification and estimation of causal effects in our setting. To illustrate potential applications, we introduce a causal Pólya urn model and demonstrate how intervention propagates effects in exchangeable data settings. Finally, we develop an algorithm that performs simultaneous causal discovery and effect estimation given multi-environment data.

PDF Details DOI

AAAI Conference 2024 Conference Paper

IT3D: Improved Text-to-3D Generation with Explicit View Synthesis

  • Yiwen Chen
  • Chi Zhang
  • Xiaofeng Yang
  • Zhongang Cai
  • Gang Yu
  • Lei Yang
  • Guosheng Lin

Recent strides in Text-to-3D techniques have been propelled by distilling knowledge from powerful large text-to-image diffusion models (LDMs). Nonetheless, existing Text-to-3D approaches often grapple with challenges such as over-saturation, inadequate detailing, and unrealistic outputs. This study presents a novel strategy that leverages explicitly synthesized multi-view images to address these issues. Our approach involves the utilization of image-to-image pipelines, empowered by LDMs, to generate posed high-quality images based on the renderings of coarse 3D models. Although the generated images mostly alleviate the aforementioned issues, challenges such as view inconsistency and significant content variance persist due to the inherent generative nature of large diffusion models, posing extensive difficulties in leveraging these images effectively. To overcome this hurdle, we advocate integrating a discriminator alongside a novel Diffusion-GAN dual training strategy to guide the training of 3D models. For the incorporated discriminator, the synthesized multi-view images are considered real data, while the renderings of the optimized 3D models function as fake data. We conduct a comprehensive set of experiments that demonstrate the effectiveness of our method over baseline approaches.

PDF Details DOI

NeurIPS Conference 2024 Conference Paper

Lever LM: Configuring In-Context Sequence to Lever Large Vision Language Models

  • Xu Yang
  • Yingzhe Peng
  • Haoxuan Ma
  • Shuo Xu
  • Chi Zhang
  • Yucheng Han
  • Hanwang Zhang

As Archimedes famously said, ``Give me a lever long enough and a fulcrum on which to place it, and I shall move the world'', in this study, we propose to use a tiny Language Model (LM), \eg, a Transformer with 67M parameters, to lever much larger Vision-Language Models (LVLMs) with 9B parameters. Specifically, we use this tiny \textbf{Lever-LM} to configure effective in-context demonstration (ICD) sequences to improve the In-Context Learinng (ICL) performance of LVLMs. Previous studies show that diverse ICD configurations like the selection and ordering of the demonstrations heavily affect the ICL performance, highlighting the significance of configuring effective ICD sequences. Motivated by this and by re-considering the the process of configuring ICD sequence, we find this is a mirror process of human sentence composition and further assume that effective ICD configurations may contain internal statistical patterns that can be captured by Lever-LM. Then a dataset with effective ICD sequences is constructed to train Lever-LM. After training, given novel queries, new ICD sequences are configured by the trained Lever-LM to solve vision-language tasks through ICL. Experiments show that these ICD sequences can improve the ICL performance of two LVLMs compared with some strong baselines in Visual Question Answering and Image Captioning, validating that Lever-LM can really capture the statistical patterns for levering LVLMs. The code is available at \url{https: //anonymous. 4open. science/r/Lever-LM-604A/}.

PDF Details DOI

IJCAI Conference 2024 Conference Paper

Reinforcement Learning for Athletic Intelligence: Lessons from the 1st “AI Olympics with RealAIGym” Competition

  • Felix Wiebe
  • Niccolò Turcato
  • Alberto Dalla Libera
  • Chi Zhang
  • Theo Vincent
  • Shubham Vyas
  • Giulio Giacomuzzo
  • Ruggero Carli

As artificial intelligence gains new capabilities, it becomes important to evaluate it on real-world tasks. In particular, the fields of robotics and reinforcement learning (RL) are lacking in standardized benchmarking tasks on real hardware. To facilitate reproducibility and stimulate algorithmic advancements, we held an AI Olympics competition at IJCAI 2023 conference based on the double pendulum system in the RealAIGym project where the participants were asked to develop a controller for the swing up and stabilization task. This paper presents the methods and results from the top participating teams and provides insights into the real-world performance of RL algorithms with respect to a baseline time-varying LQR controller.

PDF Details DOI

NeurIPS Conference 2023 Conference Paper

Active Reasoning in an Open-World Environment

  • Manjie Xu
  • Guangyuan Jiang
  • Wei Liang
  • Chi Zhang
  • Yixin Zhu

Recent advances in vision-language learning have achieved notable success on complete-information question-answering datasets through the integration of extensive world knowledge. Yet, most models operate passively, responding to questions based on pre-stored knowledge. In stark contrast, humans possess the ability to actively explore, accumulate, and reason using both newfound and existing information to tackle incomplete-information questions. In response to this gap, we introduce Conan, an interactive open-world environment devised for the assessment of active reasoning. Conan facilitates active exploration and promotes multi-round abductive inference, reminiscent of rich, open-world settings like Minecraft. Diverging from previous works that lean primarily on single-round deduction via instruction following, Conan compels agents to actively interact with their surroundings, amalgamating new evidence with prior knowledge to elucidate events from incomplete observations. Our analysis on \bench underscores the shortcomings of contemporary state-of-the-art models in active exploration and understanding complex scenarios. Additionally, we explore Abduction from Deduction, where agents harness Bayesian rules to recast the challenge of abduction as a deductive process. Through Conan, we aim to galvanize advancements in active reasoning and set the stage for the next generation of artificial intelligence agents adept at dynamically engaging in environments.

PDF Details

NeurIPS Conference 2023 Conference Paper

BeaverTails: Towards Improved Safety Alignment of LLM via a Human-Preference Dataset

  • Jiaming Ji
  • Mickel Liu
  • Josef Dai
  • Xuehai Pan
  • Chi Zhang
  • Ce Bian
  • Boyuan Chen
  • Ruiyang Sun

In this paper, we introduce the BeaverTails dataset, aimed at fostering research on safety alignment in large language models (LLMs). This dataset uniquely separates annotations of helpfulness and harmlessness for question-answering pairs, thus offering distinct perspectives on these crucial attributes. In total, we have gathered safety meta-labels for 333, 963 question-answer (QA) pairs and 361, 903 pairs of expert comparison data for both the helpfulness and harmlessness metrics. We further showcase applications of BeaverTails in content moderation and reinforcement learning with human feedback (RLHF), emphasizing its potential for practical safety measures in LLMs. We believe this dataset provides vital resources for the community, contributing towards the safe development and deployment of LLMs. Our project page is available at the following URL: https: //sites. google. com/view/pku-beavertails.

PDF Details

TIST Journal 2023 Journal Article

Configure Your Federation: Hierarchical Attention-enhanced Meta-Learning Network for Personalized Federated Learning

  • Yujia Gao
  • Pengfei Wang
  • Liang Liu
  • Chi Zhang
  • Huadong Ma

Federated learning, as a distributed machine learning framework, enables clients to conduct model training without transmitting their data to the server, which is used to solve the dilemma of data silos and data privacy. It can work well on clients having similar data characteristics and distribution. However, it has some limitations where the dataset of clients may be different in distribution, quantity, and concept in many application scenarios. Personalized federated learning is a new federated learning paradigm that aims to guarantee client personalized models’ effectiveness when collaborating with the cloud server. Intuitively, providing further facilitated collaborations for the clients with similar data characteristics and distribution can benefit personalized model building. However, due to the invisibility of client data, it is challenging to extract client characteristics and define collaborative relationships among them from a fine-grained view. Moreover, a reasonable collaborative training approach needs to be designed for a distributed server–client framework. In this article, we design a Hierarchical Attention-enhanced Meta-learning Network (HAM) to address this issue. The main advantage of HAM is that it utilizes the meta-learning approach of taking model parameters as features and learns to learn an extra model for each client to analyze similarities according to their local dataset automatically. According to its two-layers framework, HAM can reasonably achieve a tradeoff between clients’ personality and commonality and provides a hybrid model with useful information from all clients. Considering there are two networks (HAM and base network) that need to learn for each client during the federated training process, we then provide an alternative learning approach to train them in an end-to-end fashion. To further clarify the approach, we describe the personalized federated learning settings framework as FedHAM where the HAM network is distributed deployed in each client. Extensive experiments based on two datasets prove that our method outperforms state-of-the-art baselines under different evaluation metrics.

Details DOI

NeurIPS Conference 2023 Conference Paper

Constrained Policy Optimization with Explicit Behavior Density For Offline Reinforcement Learning

  • Jing Zhang
  • Chi Zhang
  • Wenjia Wang
  • Bingyi Jing

Due to the inability to interact with the environment, offline reinforcement learning (RL) methods face the challenge of estimating the Out-of-Distribution (OOD) points. Existing methods for addressing this issue either control policy to exclude the OOD action or make the $Q$ function pessimistic. However, these methods can be overly conservative or fail to identify OOD areas accurately. To overcome this problem, we propose a Constrained Policy optimization with Explicit Behavior density (CPED) method that utilizes a flow-GAN model to explicitly estimate the density of behavior policy. By estimating the explicit density, CPED can accurately identify the safe region and enable exploration within the region, resulting in less conservative learning policies. We further provide theoretical results for both the flow-GAN estimator and performance guarantee for CPED by showing that CPED can find the optimal $Q$-function value. Empirically, CPED outperforms existing alternatives on various standard offline reinforcement learning tasks, yielding higher expected returns.

PDF Details

ICRA Conference 2023 Conference Paper

Cross-Modal Monocular Localization in Prior LiDAR Maps Utilizing Semantic Consistency

  • Chi Zhang
  • Hengwang Zhao
  • Chunxiang Wang
  • Xuanlai Tang
  • Ming Yang 0002

Visual localization for mobile robots and intelligent vehicles in prior LiDAR maps can achieve high accuracy and low cost. However, algorithms for finding the cross-modal correspondences between images and LiDAR map points are not yet stable. In this paper, we propose a monocular visual localization system in prior LiDAR maps, which is based on the cross-modal registration to optimize the camera pose. To align the point clouds from vision and LiDAR map, a point-to-plane Iterative Closest Point algorithm utilizing semantic consistency is designed, and a decoupling optimization strategy is proposed to compute the affine transformation for the monocular scale ambiguity. Experiments on KITTI dataset show that utilizing the semantic consistency and geometric information of the map makes our system competitive with other methods. On the self-collected dataset, experiments on different light intensities demonstrate the robustness of the system in long-term localization tasks, and the ablation study demonstrates the effectiveness of the proposed algorithms.

Details

NeurIPS Conference 2023 Conference Paper

Evaluating and Inducing Personality in Pre-trained Language Models

  • Guangyuan Jiang
  • Manjie Xu
  • Song-Chun Zhu
  • Wenjuan Han
  • Chi Zhang
  • Yixin Zhu

Standardized and quantified evaluation of machine behaviors is a crux of understanding LLMs. In this study, we draw inspiration from psychometric studies by leveraging human personality theory as a tool for studying machine behaviors. Originating as a philosophical quest for human behaviors, the study of personality delves into how individuals differ in thinking, feeling, and behaving. Toward building and understanding human-like social machines, we are motivated to ask: Can we assess machine behaviors by leveraging human psychometric tests in a **principled** and **quantitative** manner? If so, can we induce a specific personality in LLMs? To answer these questions, we introduce the Machine Personality Inventory (MPI) tool for studying machine behaviors; MPI follows standardizedpersonality tests, built upon the Big Five Personality Factors (Big Five) theory and personality assessment inventories. By systematically evaluating LLMs with MPI, we provide the first piece of evidence demonstrating the efficacy of MPI in studying LLMs behaviors. We further devise a Personality Prompting (P$^2$) method to induce LLMs with specific personalities in a **controllable** way, capable of producing diverse and verifiable behaviors. We hope this work sheds light on future studies by adopting personality as the essential indicator for various downstream tasks, and could further motivate research into equally intriguing human-like machine behaviors.

PDF Details

YNIMG Journal 2023 Journal Article

Functional to structural plasticity in unilateral sudden sensorineural hearing loss: neuroimaging evidence

  • Yu-Ting Li
  • Ke Bai
  • Gan-Ze Li
  • Bo Hu
  • Jia-Wei Chen
  • Yu-Xuan Shang
  • Ying Yu
  • Zhu-Hong Chen

A cortical plasticity after long-duration single side deafness (SSD) is advocated with neuroimaging evidence while little is known about the short-duration SSDs. In this case-cohort study, we recruited unilateral sudden sensorineural hearing loss (SSNHL) patients and age-, gender-matched health controls (HC), followed by comprehensive neuroimaging analyses. The primary outcome measures were temporal alterations of varied dynamic functional network connectivity (dFNC) states, neurovascular coupling (NVC) and brain region volume at different stages of SSNHL. The secondary outcome measures were pure-tone audiograms of SSNHL patients before and after treatment. A total of 38 SSNHL patients (21 [55%] male; mean [standard deviation] age, 45.05 [15.83] years) and 44 HC (28 [64%] male; mean [standard deviation] age, 43.55 [12.80] years) were enrolled. SSNHL patients were categorized into subgroups based on the time from disease onset to the initial magnetic resonance imaging scan: early- (n = 16; 1-6 days), intermediate- (n = 9; 7-13 days), and late- stage (n = 13; 14-30 days) groups. We first identified slow state transitions between varied dFNC states at early-stage SSNHL, then revealed the decreased NVC restricted to the auditory cortex at the intermediate- and late-stage SSNHL. Finally, a significantly decreased volume of the left medial superior frontal gyrus (SFGmed) was observed only in the late-stage SSNHL cohort. Furthermore, the volume of the left SFGmed is robustly correlated with both disease duration and patient prognosis. Our study offered neuroimaging evidence for the evolvement from functional to structural brain alterations of SSNHL patients with disease duration less than 1 month, which may explain, from a neuroimaging perspective, why early-stage SSNHL patients have better therapeutic responses and hearing recovery.

Details DOI

NeurIPS Conference 2023 Conference Paper

Interactive Visual Reasoning under Uncertainty

  • Manjie Xu
  • Guangyuan Jiang
  • Wei Liang
  • Chi Zhang
  • Yixin Zhu

One of the fundamental cognitive abilities of humans is to quickly resolve uncertainty by generating hypotheses and testing them via active trials. Encountering a novel phenomenon accompanied by ambiguous cause-effect relationships, humans make hypotheses against data, conduct inferences from observation, test their theory via experimentation, and correct the proposition if inconsistency arises. These iterative processes persist until the underlying mechanism becomes clear. In this work, we devise the IVRE (pronounced as "ivory" ) environment for evaluating artificial agents' reasoning ability under uncertainty. IVRE is an interactive environment featuring rich scenarios centered around Blicket detection. Agents in IVRE are placed into environments with various ambiguous action-effect pairs and asked to determine each object's role. They are encouraged to propose effective and efficient experiments to validate their hypotheses based on observations and actively gather new information. The game ends when all uncertainties are resolved or the maximum number of trials is consumed. By evaluating modern artificial agents in IVRE, we notice a clear failure of today's learning methods compared to humans. Such inefficacy in interactive reasoning ability under uncertainty calls for future research in building human-like intelligence.

PDF Details

EAAI Journal 2023 Journal Article

Nested physics-informed neural network for analysis of transient flows in natural gas pipelines

  • Chi Zhang
  • Abdollah Shafieezadeh

Natural gas pipeline systems are commonly designed under the assumption of constant supply and demand flow conditions. This is while gas flows are transient because of the compression stations, presence of gas storage facilities and fluctuating supply and consumer demands. Analyzing such transient flows substantially benefits design, control, and monitoring of natural gas pipeline systems. The nonlinear partial differential equations describing the physics of transient flows in pipelines are solved using conventional methods, which are computationally demanding especially for uncertainty quantification purposes where many simulations are required. In this study, we propose an alternative physics-informed neural network (PINN) framework for the transient analysis of pipeline networks that can perform transient flow analysis in natural gas pipelines that the original PINNs cannot solve due to the high complexity of the problem. We propose a nested structure for the PINNs with a loss model that greatly reduces the number of tasks in the emergent complex multi-task learning process. We also integrate an adaptive weights approach that tackles the imbalanced gradients caused by the extremely large coefficients in the equations of the natural gas pipeline problem. The proposed framework, for the first time, can produce accurate results for the complex natural gas pipeline network problem using PINNs. Furthermore, we investigate the parameterization of the nested PINNs as a surrogate model for the natural gas pipeline system. With merely 26% more training costs, the surrogate model can perform the transient flow analysis given thousands of realizations of parameters in a split millisecond, while the costs of performing the same simulations using conventional methods can be prohibitively expensive. This can greatly boost the efficiency of complex many-query analyses such as sensitivity analysis, uncertainty propagation and design optimization.

Details DOI

ICLR Conference 2023 Conference Paper

On the Perils of Cascading Robust Classifiers

  • Ravi Mangal
  • Zifan Wang 0001
  • Chi Zhang
  • Klas Leino
  • Corina S. Pasareanu
  • Matt Fredrikson

Ensembling certifiably robust neural networks is a promising approach for improving the \emph{certified robust accuracy} of neural models. Black-box ensembles that assume only query-access to the constituent models (and their robustness certifiers) during prediction are particularly attractive due to their modular structure. Cascading ensembles are a popular instance of black-box ensembles that appear to improve certified robust accuracies in practice. However, we show that the robustness certifier used by a cascading ensemble is unsound. That is, when a cascading ensemble is certified as locally robust at an input $x$ (with respect to $\epsilon$), there can be inputs $x'$ in the $\epsilon$-ball centered at $x$, such that the cascade's prediction at $x'$ is different from $x$ and thus the ensemble is not locally robust. Our theoretical findings are accompanied by empirical results that further demonstrate this unsoundness. We present a new attack against cascading ensembles and show that: (1) there exists an adversarial input for up to 88\% of the samples where the ensemble claims to be certifiably robust and accurate; and (2) the accuracy of a cascading ensemble under our attack is as low as 11\% when it claims to be certifiably robust and accurate on 97\% of the test set. Our work reveals a critical pitfall of cascading certifiably robust models by showing that the seemingly beneficial strategy of cascading can actually hurt the robustness of the resulting ensemble. Our code is available at https://github.com/TristaChi/ensembleKW.

Details

AAAI Conference 2023 Conference Paper

Reachability Analysis of Neural Network Control Systems

  • Chi Zhang
  • Wenjie Ruan
  • Peipei Xu

Neural network controllers (NNCs) have shown great promise in autonomous and cyber-physical systems. Despite the various verification approaches for neural networks, the safety analysis of NNCs remains an open problem. Existing verification approaches for neural network control systems (NNCSs) either can only work on a limited type of activation functions, or result in non-trivial over-approximation errors with time evolving. This paper proposes a verification framework for NNCS based on Lipschitzian optimisation, called DeepNNC. We first prove the Lipschitz continuity of closed-loop NNCSs by unrolling and eliminating the loops. We then reveal the working principles of applying Lipschitzian optimisation on NNCS verification and illustrate it by verifying an adaptive cruise control model. Compared to state-of-the-art verification approaches, DeepNNC shows superior performance in terms of efficiency and accuracy over a wide range of NNCs. We also provide a case study to demonstrate the capability of DeepNNC to handle a real-world, practical, and complex system. Our tool DeepNNC is available at https://github.com/TrustAI/DeepNNC.

PDF Details DOI

NeurIPS Conference 2023 Conference Paper

SustainGym: Reinforcement Learning Environments for Sustainable Energy Systems

  • Christopher Yeh
  • Victor Li
  • Rajeev Datta
  • Julio Arroyo
  • Nicolas Christianson
  • Chi Zhang
  • Yize Chen
  • Mohammad Mehdi Hosseini

The lack of standardized benchmarks for reinforcement learning (RL) in sustainability applications has made it difficult to both track progress on specific domains and identify bottlenecks for researchers to focus their efforts. In this paper, we present SustainGym, a suite of five environments designed to test the performance of RL algorithms on realistic sustainable energy system tasks, ranging from electric vehicle charging to carbon-aware data center job scheduling. The environments test RL algorithms under realistic distribution shifts as well as in multi-agent settings. We show that standard off-the-shelf RL algorithms leave significant room for improving performance and highlight the challenges ahead for introducing RL to real-world sustainability tasks.

PDF Details

NeurIPS Conference 2022 Conference Paper

Causal Inference with Non-IID Data using Linear Graphical Models

  • Chi Zhang
  • Karthika Mohan
  • Judea Pearl

Traditional causal inference techniques assume data are independent and identically distributed (IID) and thus ignores interactions among units. However, a unit’s treatment may affect another unit's outcome (interference), a unit’s treatment may be correlated with another unit’s outcome, or a unit’s treatment and outcome may be spuriously correlated through another unit. To capture such nuances, we model the data generating process using causal graphs and conduct a systematic analysis of the bias caused by different types of interactions when computing causal effects. We derive theorems to detect and quantify the interaction bias, and derive conditions under which it is safe to ignore interactions. Put differently, we present conditions under which causal effects can be computed with negligible bias by assuming that samples are IID. Furthermore, we develop a method to eliminate bias in cases where blindly assuming IID is expected to yield a significantly biased estimate. Finally, we test the coverage and performance of our methods through simulations.

PDF Details

TMLR Journal 2022 Journal Article

Degradation Attacks on Certifiably Robust Neural Networks

  • Klas Leino
  • Chi Zhang
  • Ravi Mangal
  • Matt Fredrikson
  • Bryan Parno
  • Corina Pasareanu

Certifiably robust neural networks protect against adversarial examples by employing run-time defenses that check if the model is certifiably locally robust at the input under evaluation. We show through examples and experiments that any defense (whether complete or incomplete) based on checking local robustness is inherently over-cautious. Specifically, such defenses flag inputs for which local robustness checks fail, but yet that are not adversarial; i.e., they are classified consistently with all valid inputs within a distance of $\epsilon$. As a result, while a norm-bounded adversary cannot change the classification of an input, it can use norm-bounded changes to degrade the utility of certifiably robust networks by forcing them to reject otherwise correctly classifiable inputs. We empirically demonstrate the efficacy of such attacks against state-of-the-art certifiable defenses. Our code is available at https://github.com/ravimangal/degradation-attacks.

PDF Details

IROS Conference 2022 Conference Paper

Grasping State Analysis of Soft Manipulator Based on Flexible Tactile Sensor Array

  • Haoyuan Wang
  • Hongge Ru
  • Hongliang Lei
  • Chi Zhang
  • Cheng Han
  • Hao Wu 0028
  • Jian Huang 0001

Although the grasping state analysis is vital in the study of manipulators, the grasping state analysis of soft manipulators as an independent research topic is not much so far. This paper proposes a novel pneumatic soft manipulator with a flexible tactile sensor array (SM-FTSA). The flexible tactile sensor array comprises piezoresistive materials with a porous structure. An equal potential approach is adopted to realize the collection of tactile signals of the SM-FTSA. Inspired by the grasping analysis of rigid manipulators, we propose 4 grasping states for the SM-FTSA, including inflating, shaking, stable, and slipping. Based on the experimental data, we conduct grasping experiments on 12 objects with SM-FTSA, and we propose 10 features that reflect the grasping state. Several machine learning methods are utilized to classify the grasping state. Among them, the Random Forest method presents the best performance, and the average classification accuracy reaches 99%.

Details

NeurIPS Conference 2022 Conference Paper

Hierarchical Normalization for Robust Monocular Depth Estimation

  • Chi Zhang
  • Wei Yin
  • Billzb Wang
  • Gang Yu
  • Bin Fu
  • Chunhua Shen

In this paper, we address monocular depth estimation with deep neural networks. To enable training of deep monocular estimation models with various sources of datasets, state-of-the-art methods adopt image-level normalization strategies to generate affine-invariant depth representations. However, learning with the image-level normalization mainly emphasizes the relations of pixel representations with the global statistic in the images, such as the structure of the scene, while the fine-grained depth difference may be overlooked. In this paper, we propose a novel multi-scale depth normalization method that hierarchically normalizes the depth representations based on spatial information and depth distributions. Compared with previous normalization strategies applied only at the holistic image level, the proposed hierarchical normalization can effectively preserve the fine-grained details and improve accuracy. We present two strategies that define the hierarchical normalization contexts in the depth domain and the spatial domain, respectively. Our extensive experiments show that the proposed normalization strategy remarkably outperforms previous normalization methods, and we set new state-of-the-art on five zero-shot transfer benchmark datasets.

PDF Details

AAMAS Conference 2022 Conference Paper

Intelligent Communication over Realistic Wireless Networks in Multi-Agent Cooperative Games

  • Diyi Hu
  • Chi Zhang
  • Viktor Prasanna
  • Bhaskar Krishnamachari

In MARL, communication among agents is essential to establish cooperation. Over the realistic wireless network, many factors can affect transmission reliability, especially considering that the wireless network condition varies with agents’ mobility. We propose a framework that improves the intelligence of communication over realistic wireless networks in two fundamental aspects: (1) When: Agents learn the timing of communication based on message importance and wireless channel condition. We further propose a communication lagging technique to make the training end-to-end differentiable. (2) What: Agents augment message contents with wireless network measurements. The messages improve both the game and communication actions of the agents. Experiments on a standard environment show that compared with state-of-the-art, our framework enables more intelligent collaboration and thus achieves significantly better game performance, convergence speed and communication efficiency.

PDF

AAAI Conference 2022 Conference Paper

Multi-Centroid Representation Network for Domain Adaptive Person Re-ID

  • Yuhang Wu
  • Tengteng Huang
  • Haotian Yao
  • Chi Zhang
  • Yuanjie Shao
  • Chuchu Han
  • Changxin Gao
  • Nong Sang

Recently, many approaches tackle the Unsupervised Domain Adaptive person re-identification (UDA re-ID) problem through pseudo-label-based contrastive learning. During training, a uni-centroid representation is obtained by simply averaging all the instance features from a cluster with the same pseudo label. However, a cluster may contain images with different identities (label noises) due to the imperfect clustering results, which makes the uni-centroid representation inappropriate. In this paper, we present a novel Multi-Centroid Memory (MCM) to adaptively capture different identity information within the cluster. MCM can effectively alleviate the issue of label noises by selecting proper positive/negative centroids for the query image. Moreover, we further propose two strategies to improve the contrastive learning process. First, we present a Domain-Specific Contrastive Learning (DSCL) mechanism to fully explore intradomain information by comparing samples only from the same domain. Second, we propose Second-Order Nearest Interpolation (SONI) to obtain abundant and informative negative samples. We integrate MCM, DSCL, and SONI into a unified framework named Multi-Centroid Representation Network (MCRN). Extensive experiments demonstrate the superiority of MCRN over state-of-the-art approaches on multiple UDA re-ID tasks and fully unsupervised re-ID tasks.

PDF Details

NeurIPS Conference 2022 Conference Paper

On the Learning Mechanisms in Physical Reasoning

  • Shiqian Li
  • Kewen Wu
  • Chi Zhang
  • Yixin Zhu

Is dynamics prediction indispensable for physical reasoning? If so, what kind of roles do the dynamics prediction modules play during the physical reasoning process? Most studies focus on designing dynamics prediction networks and treating physical reasoning as a downstream task without investigating the questions above, taking for granted that the designed dynamics prediction would undoubtedly help the reasoning process. In this work, we take a closer look at this assumption, exploring this fundamental hypothesis by comparing two learning mechanisms: Learning from Dynamics (LfD) and Learning from Intuition (LfI). In the first experiment, we directly examine and compare these two mechanisms. Results show a surprising finding: Simple LfI is better than or on par with state-of-the-art LfD. This observation leads to the second experiment with Ground-truth Dynamics (GD), the ideal case of LfD wherein dynamics are obtained directly from a simulator. Results show that dynamics, if directly given instead of approximated, would achieve much higher performance than LfI alone on physical reasoning; this essentially serves as the performance upper bound. Yet practically, LfD mechanism can only predict Approximate Dynamics (AD) using dynamics learning modules that mimic the physical laws, making the following downstream physical reasoning modules degenerate into the LfI paradigm; see the third experiment. We note that this issue is hard to mitigate, as dynamics prediction errors inevitably accumulate in the long horizon. Finally, in the fourth experiment, we note that LfI, the extremely simpler strategy when done right, is more effective in learning to solve physical reasoning problems. Taken together, the results on the challenging benchmark of PHYRE show that LfI is, if not better, as good as LfD with bells and whistles for dynamics prediction. However, the potential improvement from LfD, though challenging, remains lucrative.

PDF Details

YNIMG Journal 2022 Journal Article

Residual RAKI: A hybrid linear and non-linear approach for scan-specific k-space deep learning

  • Chi Zhang
  • Steen Moeller
  • Omer Burak Demirel
  • Kâmil Uğurbil
  • Mehmet Akçakaya

Parallel imaging is the most clinically used acceleration technique for magnetic resonance imaging (MRI) in part due to its easy inclusion into routine acquisitions. In k-space based parallel imaging reconstruction, sub-sampled k-space data are interpolated using linear convolutions. At high acceleration rates these methods have inherent noise amplification and reduced image quality. On the other hand, non-linear deep learning methods provide improved image quality at high acceleration, but the availability of training databases for different scans, as well as their interpretability hinder their adaptation. In this work, we present an extension of Robust Artificial-neural-networks for k-space Interpolation (RAKI), called residual-RAKI (rRAKI), which achieves scan-specific machine learning reconstruction using a hybrid linear and non-linear methodology. In rRAKI, non-linear CNNs are trained jointly with a linear convolution implemented via a skip connection. In effect, the linear part provides a baseline reconstruction, while the non-linear CNN that runs in parallel provides further reduction of artifacts and noise arising from the linear part. The explicit split between the linear and non-linear aspects of the reconstruction also help improve interpretability compared to purely non-linear methods. Experiments were conducted on the publicly available fastMRI datasets, as well as high-resolution anatomical imaging, comparing GRAPPA and its variants, compressed sensing, RAKI, Scan Specific Artifact Reduction in K-space (SPARK) and the proposed rRAKI. Additionally, highly-accelerated simultaneous multi-slice (SMS) functional MRI reconstructions were also performed, where the proposed rRAKI was compred to Read-out SENSE-GRAPPA and RAKI. Our results show that the proposed rRAKI method substantially improves the image quality compared to conventional parallel imaging, and offers sharper images compared to SPARK and ℓ1-SPIRiT. Furthermore, rRAKI shows improved preservation of time-varying dynamics compared to both parallel imaging and RAKI in highly-accelerated SMS fMRI.

Details DOI

AIJ Journal 2021 Journal Article

Distributed optimization for degenerate loss functions arising from over-parameterization

  • Chi Zhang
  • Qianxiao Li

We consider distributed optimization with degenerate loss functions, where the optimal sets of local loss functions have a non-empty intersection. This regime often arises in optimizing large-scale multi-agent AI systems (e. g. , deep learning systems), where the number of trainable weights far exceeds the number of training samples, leading to highly degenerate loss surfaces. Under appropriate conditions, we prove that distributed gradient descent in this case converges even when communication is arbitrarily less frequent, which is not the case for non-degenerate loss functions. Moreover, we quantitatively analyze the convergence rate, as well as the communication and computation trade-off, providing insights into designing efficient distributed optimization algorithms. Our theoretical findings are confirmed by both distributed convex optimization and deep learning experiments.

Details DOI

NeurIPS Conference 2021 Conference Paper

Spatial Ensemble: a Novel Model Smoothing Mechanism for Student-Teacher Framework

  • Tengteng Huang
  • Yifan Sun
  • Xun Wang
  • Haotian Yao
  • Chi Zhang

Model smoothing is of central importance for obtaining a reliable teacher model in the student-teacher framework, where the teacher generates surrogate supervision signals to train the student. A popular model smoothing method is the Temporal Moving Average (TMA), which continuously averages the teacher parameters with the up-to-date student parameters. In this paper, we propose ''Spatial Ensemble'', a novel model smoothing mechanism in parallel with TMA. Spatial Ensemble randomly picks up a small fragment of the student model to directly replace the corresponding fragment of the teacher model. Consequentially, it stitches different fragments of historical student models into a unity, yielding the ''Spatial Ensemble'' effect. Spatial Ensemble obtains comparable student-teacher learning performance by itself and demonstrates valuable complementarity with temporal moving average. Their integration, named Spatial-Temporal Smoothing, brings general (sometimes significant) improvement to the student-teacher learning framework on a variety of state-of-the-art methods. For example, based on the self-supervised method BYOL, it yields +0. 9% top-1 accuracy improvement on ImageNet, while based on the semi-supervised approach FixMatch, it increases the top-1 accuracy by around +6% on CIFAR-10 when only few training labels are available. Codes and models are available at: https: //github. com/tengteng95/Spatial_Ensemble.

PDF Details

AAAI Conference 2020 Conference Paper

A Simultaneous Discover-Identify Approach to Causal Inference in Linear Models

  • Chi Zhang
  • Bryant Chen
  • Judea Pearl

Modern causal analysis involves two major tasks, discovery and identification. The first aims to learn a causal structure compatible with the available data, the second leverages that structure to estimate causal effects. Rather than performing the two tasks in tandem, as is usually done in the literature, we propose a symbiotic approach in which the two are performed simultaneously for mutual benefit; information gained through identification helps causal discovery and vice versa. This approach enables the usage of Verma constraints, which remain dormant in constraint-based methods of discovery, and permit us to learn more complete structures, hence identify a larger set of causal effects than previously achievable with standard methods.

PDF Details

AAAI Conference 2020 Conference Paper

Fast and Efficient Boolean Matrix Factorization by Geometric Segmentation

  • Changlin Wan
  • Wennan Chang
  • Tong Zhao
  • Mengya Li
  • Sha Cao
  • Chi Zhang

Boolean matrix has been used to represent digital information in many fields, including bank transaction, crime records, natural language processing, protein-protein interaction, etc. Boolean matrix factorization (BMF) aims to find an approximation of a binary matrix as the Boolean product of two low rank Boolean matrices, which could generate vast amount of information for the patterns of relationships between the features and samples. Inspired by binary matrix permutation theories and geometric segmentation, we developed a fast and efficient BMF approach, called MEBF (Median Expansion for Boolean Factorization). Overall, MEBF adopted a heuristic approach to locate binary patterns presented as submatrices that are dense in 1’s. At each iteration, MEBF permutates the rows and columns such that the permutated matrix is approximately Upper Triangular-Like (UTL) with socalled Simultaneous Consecutive-ones Property (SC1P). The largest submatrix dense in 1 would lie on the upper triangular area of the permutated matrix, and its location was determined based on a geometric segmentation of a triangular. We compared MEBF with other state of the art approaches on data scenarios with different density and noise levels. MEBF demonstrated superior performances in lower reconstruction error, and higher computational efficiency, as well as more accurate density patterns than popular methods such as ASSO, PANDA and Message Passing. We demonstrated the application of MEBF on both binary and non-binary data sets, and revealed its further potential in knowledge retrieving and data denoising.

PDF Details

EAAI Journal 2020 Journal Article

Finding the k shortest paths by ripple-spreading algorithms

  • Xiao-Bing Hu
  • Chi Zhang
  • Gong-Peng Zhang
  • Ming-Kong Zhang
  • Hang Li
  • Mark S. Leeson
  • Jian-Qin Liao

The k shortest paths problem ( k -SPP) is fundamentally important to both theoretical and application researches on computational intelligence. Inspired by the natural ripple-spreading phenomenon that occurs on a water surface, this paper proposes a novel ripple-spreading algorithm (RSA). RSA differs from many existing methods which need to reconstruct route networks or to sweep the network for k times, and it can identify the k shortest paths by a single run of ripple relay race in the original route network. Besides the k -SPP in normal route networks, the RSA can also be extended, without losing optimality and effectiveness, to some time-window networks (where various waiting behaviors at nodes are introduced) and dynamical networks (where the network topology and link costs may change over time due to factors such as moving obstacles and spreading disasters). For one-to-all k -SPP, which aims to find all the k shortest paths from a given source to every other node in a network (no matter with or without time windows at nodes, and no matter whether the network topology and link costs can change over time or not), the RSA can still find out all required solutions using only a single run, while the computational complexity is exactly the same as that for the one-to-one k -SPP, i. e. , O ( k × N L × N A T U ), where N L is the number of links in the network, and N A T U is the average simulated time units for a ripple to travel through a link. The comparative experimental results illustrate the effectiveness and efficiency of the proposed RSA.

Details DOI

NeurIPS Conference 2020 Conference Paper

Geometric All-way Boolean Tensor Decomposition

  • Changlin Wan
  • Wennan Chang
  • Tong Zhao
  • Sha Cao
  • Chi Zhang

Boolean tensor has been broadly utilized in representing high dimensional logical data collected on spatial, temporal and/or other relational domains. Boolean Tensor Decomposition (BTD) factorizes a binary tensor into the Boolean sum of multiple rank-1 tensors, which is an NP-hard problem. Existing BTD methods have been limited by their high computational cost, in applications to large scale or higher order tensors. In this work, we presented a computationally efficient BTD algorithm, namely Geometric Expansion for all-order Tensor Factorization (GETF), that sequentially identifies the rank-1 basis components for a tensor from a geometric perspective. We conducted rigorous theoretical analysis on the validity as well as algorithemic efficiency of GETF in decomposing all-order tensor. Experiments on both synthetic and real-world data demonstrated that GETF has significantly improved performance in reconstruction accuracy, extraction of latent structures and it is an order of magnitude faster than other state-of-the-art methods.

PDF Details

IS Journal 2020 Journal Article

Joint Intelligence Ranking by Federated Multiplicative Update

  • Chi Zhang
  • Yu Liu
  • Le Wang
  • Yuehu Liu
  • Li Li
  • Nanning Zheng

The joint intelligence ranking of intelligent systems like autonomous driving is of great importance for building a more general, extensive, and universally accepted intelligence evaluation scheme. However, due to issues such as privacy security and industry or area competition, the integration of isolated test results may face large unimaginable difficulty in information security and encrypted model training. To address this, we derive the federated multiplicative update (FMU) algorithm with boundary constraints to solve the nonnegative matrix factorization based joint intelligence ranking. The encrypted learning process is developed to alternate original computation steps in multiplicative update algorithms. Owning feasible property for the fast convergence and secure exchange of variables, the proposed framework outperforms the previous work on both real and simulated data. Further experimental analysis reveals that the introduced federated mechanism does not harm the overall time efficiency.

Details DOI

NeurIPS Conference 2020 Conference Paper

Learning Disentangled Representations of Videos with Missing Data

  • Armand Comas
  • Chi Zhang
  • Zlatan Feric
  • Octavia Camps
  • Rose Yu

Missing data poses significant challenges while learning representations of video sequences. We present Disentangled Imputed Video autoEncoder (DIVE), a deep generative model that imputes and predicts future video frames in the presence of missing data. Specifically, DIVE introduces a missingness latent variable, disentangles the hidden video representations into static and dynamic appearance, pose, and missingness factors for each object, while it imputes each object trajectory where data is missing. On a moving MNIST dataset with various missing scenarios, DIVE outperforms the state of the art baselines by a substantial margin. We also present comparisons on a real-world MOTSChallenge pedestrian dataset, which demonstrates the practical value of our method in a more realistic setting. Our code can be found in https: //github. com/Rose-STL-Lab/DIVE.

PDF Details

AAAI Conference 2020 Conference Paper

Machine Number Sense: A Dataset of Visual Arithmetic Problems for Abstract and Relational Reasoning

  • Wenhe Zhang
  • Chi Zhang
  • Yixin Zhu
  • Song-Chun Zhu

As a comprehensive indicator of mathematical thinking and intelligence, the number sense (Dehaene 2011) bridges the induction of symbolic concepts and the competence of problemsolving. To endow such a crucial cognitive ability to machine intelligence, we propose a dataset, Machine Number Sense (MNS), consisting of visual arithmetic problems automatically generated using a grammar model—And-Or Graph (AOG). These visual arithmetic problems are in the form of geometric figures: each problem has a set of geometric shapes as its context and embedded number symbols. Solving such problems is not trivial; the machine not only has to recognize the number, but also to interpret the number with its contexts, shapes, and relations (e. g. , symmetry) together with proper operations. We benchmark the MNS dataset using four predominant neural network models as baselines in this visual reasoning task. Comprehensive experiments show that current neural-network-based models still struggle to understand number concepts and relational operations. We show that a simple brute-force search algorithm could work out some of the problems without context information. Crucially, taking geometric context into account by an additional perception module would provide a sharp performance gain with fewer search steps. Altogether, we call for attention in fusing the classic search-based algorithms with modern neural networks to discover the essential number concepts in future research.

PDF Details

IJCAI Conference 2019 Conference Paper

Decentralized Optimization with Edge Sampling

  • Chi Zhang
  • Qianxiao Li
  • Peilin Zhao

In this paper, we propose a decentralized distributed algorithm with stochastic communication among nodes, building on a sampling method called "edge sampling''. Such a sampling algorithm allows us to avoid the heavy peer-to-peer communication cost when combining neighboring weights on dense networks while still maintains a comparable convergence rate. In particular, we quantitatively analyze its theoretical convergence properties, as well as the optimal sampling rate over the underlying network. When compared with previous methods, our solution is shown to be unbiased, communication-efficient and suffers from lower sampling variances. These theoretical findings are validated by both numerical experiments on the mixing rates of Markov Chains and distributed machine learning problems.

PDF Details

NeurIPS Conference 2019 Conference Paper

Learning Perceptual Inference by Contrasting

  • Chi Zhang
  • Baoxiong Jia
  • Feng Gao
  • Yixin Zhu
  • Hongjing Lu
  • Song-Chun Zhu

“Thinking in pictures, ” [1] i. e. , spatial-temporal reasoning, effortless and instantaneous for humans, is believed to be a significant ability to perform logical induction and a crucial factor in the intellectual history of technology development. Modern Artificial Intelligence (AI), fueled by massive datasets, deeper models, and mighty computation, has come to a stage where (super-)human-level performances are observed in certain specific tasks. However, current AI's ability in “thinking in pictures” is still far lacking behind. In this work, we study how to improve machines' reasoning ability on one challenging task of this kind: Raven's Progressive Matrices (RPM). Specifically, we borrow the very idea of “contrast effects” from the field of psychology, cognition, and education to design and train a permutation-invariant model. Inspired by cognitive studies, we equip our model with a simple inference module that is jointly trained with the perception backbone. Combining all the elements, we propose the Contrastive Perceptual Inference network (CoPINet) and empirically demonstrate that CoPINet sets the new state-of-the-art for permutation-invariant models on two major datasets. We conclude that spatial-temporal reasoning depends on envisaging the possibilities consistent with the relations between objects and can be solved from pixel-level inputs.

PDF Details

IROS Conference 2019 Conference Paper

Learning Safe Unlabeled Multi-Robot Planning with Motion Constraints

  • Arbaaz Khan
  • Chi Zhang
  • Shuo Li
  • Jiayue Wu
  • Brent Schlotfeldt
  • Sarah Y. Tang
  • Alejandro Ribeiro
  • Osbert Bastani

In this paper, we present a learning approach to goal assignment and trajectory planning for unlabeled robots operating in 2D, obstacle-filled workspaces. More specifically, we tackle the unlabeled multi-robot motion planning problem with motion constraints as a multi-agent reinforcement learning problem with some sparse global reward. In contrast with previous works, which formulate an entirely new hand-crafted optimization cost or trajectory generation algorithm for a different robot dynamic model, our framework is a general approach that is applicable to arbitrary robot models. Further, by using the velocity obstacle, we devise a smooth projection that guarantees collision free trajectories for all robots with respect to their neighbors and obstacles. The efficacy of our algorithm is demonstrated through varied simulations. A video describing our method and results can be found here.

Details

AAAI Conference 2019 Conference Paper

MetaStyle: Three-Way Trade-off among Speed, Flexibility, and Quality in Neural Style Transfer

  • Chi Zhang
  • Yixin Zhu
  • Song-Chun Zhu

An unprecedented booming has been witnessed in the research area of artistic style transfer ever since Gatys et al. introduced the neural method. One of the remaining challenges is to balance a trade-off among three critical aspects—speed, flexibility, and quality: (i) the vanilla optimization-based algorithm produces impressive results for arbitrary styles, but is unsatisfyingly slow due to its iterative nature, (ii) the fast approximation methods based on feed-forward neural networks generate satisfactory artistic effects but bound to only a limited number of styles, and (iii) feature-matching methods like AdaIN achieve arbitrary style transfer in a real-time manner but at a cost of the compromised quality. We find it considerably difficult to balance the trade-off well merely using a single feed-forward step and ask, instead, whether there exists an algorithm that could adapt quickly to any style, while the adapted model maintains high efficiency and good image quality. Motivated by this idea, we propose a novel method, coined MetaStyle, which formulates the neural style transfer as a bilevel optimization problem and combines learning with only a few post-processing update steps to adapt to a fast approximation model with satisfying artistic effects, comparable to the optimization-based methods for an arbitrary style. The qualitative and quantitative analysis in the experiments demonstrates that the proposed approach achieves high-quality arbitrary artistic style transfer effectively, with a good trade-off among speed, flexibility, and quality.

PDF Details

AAAI Conference 2019 Conference Paper

Mirroring without Overimitation: Learning Functionally Equivalent Manipulation Actions

  • Hangxin Liu
  • Chi Zhang
  • Yixin Zhu
  • Chenfanfu Jiang
  • Song-Chun Zhu

This paper presents a mirroring approach, inspired by the neuroscience discovery of the mirror neurons, to transfer demonstrated manipulation actions to robots. Designed to address the different embodiments between a human (demonstrator) and a robot, this approach extends the classic robot Learning from Demonstration (LfD) in the following aspects: i) It incorporates fine-grained hand forces collected by a tactile glove in demonstration to learn robot’s fine manipulative actions; ii) Through model-free reinforcement learning and grammar induction, the demonstration is represented by a goal-oriented grammar consisting of goal states and the corresponding forces to reach the states, independent of robot embodiments; iii) A physics-based simulation engine is applied to emulate various robot actions and mirrors the actions that are functionally equivalent to the human’s in the sense of causing the same state changes by exerting similar forces. Through this approach, a robot reasons about which forces to exert and what goals to achieve to generate actions (i. e. , mirroring), rather than strictly mimicking demonstration (i. e. , overimitation). Thus the embodiment difference between a human and a robot is naturally overcome. In the experiment, we demonstrate the proposed approach by teaching a real Baxter robot with a complex manipulation task involving haptic feedback—opening medicine bottles.

PDF Details

IJCAI Conference 2018 Conference Paper

Learning Unmanned Aerial Vehicle Control for Autonomous Target Following

  • Siyi Li
  • Tianbo Liu
  • Chi Zhang
  • Dit-Yan Yeung
  • Shaojie Shen

While deep reinforcement learning (RL) methods have achieved unprecedented successes in a range of challenging problems, their applicability has been mainly limited to simulation or game domains due to the high sample complexity of the trial-and-error learning process. However, real-world robotic applications often need a data-efficient learning process with safety-critical constraints. In this paper, we consider the challenging problem of learning unmanned aerial vehicle (UAV) control for tracking a moving target. To acquire a strategy that combines perception and control, we represent the policy by a convolutional neural network. We develop a hierarchical approach that combines a model-free policy gradient method with a conventional feedback proportional-integral-derivative (PID) controller to enable stable learning without catastrophic failure. The neural network is trained by a combination of supervised learning from raw images and reinforcement learning from games of self-play. We show that the proposed approach can learn a target following policy in a simulator efficiently and the learned behavior can be successfully transferred to the DJI quadrotor platform for real-world UAV control.

PDF Details

AAAI Conference 2017 Conference Paper

Deep Manifold Learning of Symmetric Positive Definite Matrices with Application to Face Recognition

  • Zhen Dong
  • Su Jia
  • Chi Zhang
  • Mingtao Pei
  • Yuwei Wu

In this paper, we aim to construct a deep neural network which embeds high dimensional symmetric positive definite (SPD) matrices into a more discriminative low dimensional SPD manifold. To this end, we develop two types of basic layers: a 2D fully connected layer which reduces the dimensionality of the SPD matrices, and a symmetrically clean layer which achieves non-linear mapping. Specifically, we extend the classical fully connected layer such that it is suitable for SPD matrices, and we further show that SPD matrices with symmetric pair elements setting zero operations are still symmetric positive definite. Finally, we complete the construction of the deep neural network for SPD manifold learning by stacking the two layers. Experiments on several face datasets demonstrate the effectiveness of the proposed method.

PDF Details

ICRA Conference 2015 Conference Paper

Adaptive human-centered representation for activity recognition of multiple individuals from 3D point cloud sequences

  • Hao Zhang 0011
  • Christopher M. Reardon
  • Chi Zhang
  • Lynne E. Parker

Activity recognition of multi-individuals (ARMI) within a group, which is essential to practical human-centered robotics applications such as childhood education, is a particularly challenging and previously not well studied problem. We present a novel adaptive human-centered (AdHuC) representation based on local spatio-temporal features (LST) to address ARMI in a sequence of 3D point clouds. Our human-centered detector constructs affiliation regions to associate LST features with humans by mining depth data and using a cascade of rejectors to localize humans in 3D space. Then, features are detected within each affiliation region, which avoids extracting irrelevant features from dynamic background clutter and addresses moving cameras on mobile robots. Our feature descriptor is able to adapt its support region to linear perspective view variations and encode multi-channel information (i. e. , color and depth) to construct the final representation. Empirical studies validate that the AdHuC representation obtains promising performance on ARMI using a Meka humanoid robot to play multi-people Simon Says games. Experiments on benchmark datasets further demonstrate that our adaptive human-centered representation outperforms previous approaches for activity recognition from color-depth data.

Details

IROS Conference 2015 Conference Paper

Feature Space Decomposition for effective robot adaptation

  • Chi Zhang
  • Hao Zhang 0011
  • Lynne E. Parker

Adaptation is an essential capability for intelligent robots to work in new environments. In the learning framework of Programming by Demonstration (PbD) and Reinforcement Learning (RL), a robot usually learns skills from a latent feature space obtained by dimension reduction techniques. Because the latent space is optimized for a specific environment during the training phase, it typically contains fewer variations. Accordingly, searching for a solution within the latent space can be less effective for robot adaptation to new environments with unseen changes. In this paper, we propose a novel Feature Space Decomposition (FSD) approach to effectively address the robot adaptation problem, which is directly applicable to the learning framework based on PbD and RL. Our FSD method decomposes the high-dimensional original features extracted from the demonstration data into principal and non-principal feature space. Then, the non-principal features are used to form a new low-dimensional search space for autonomous robot adaptation based on RL, which is initialized using a generalized trajectory represented by a Gaussian Mixture Model that is learned from the principal features. The scalability of our FSD approach guarantees that optimal solutions can be found in the new non-principal space, if they exist in the original feature space. Experimental results on real robots validate that our FSD approach enables the robots to effectively adapt to new environments, and is usually able to find optimal solutions more quickly than traditional approaches when significant environment changes occur.

Details

ICRA Conference 1996 Conference Paper

Vision-based robot localization without explicit object models

  • Gregory Dudek
  • Chi Zhang

We consider the problem of locating a robot in an initially-unfamiliar environment from visual input. The robot is not given a map of the environment, but it does have access to a collection of training examples, each of which specifies the video image observed when the robot is at a particular location and orientation. We address two variants of this problem: how to estimate translation of a moving robot assuming the orientation is known, and how to estimate translation and orientation for a mobile robot. Performing scene reconstruction to construct a metric map of the environment using only video images is difficult. We avoid this by using an approach in which the robot learns to convert a set of image measurements into a representation of its pose (position and orientation). This provides a metric estimate of the robot's location within a region covered by the statistical map we build. Localization can be performed online without a prior location estimate, The conversion from visual data to camera pose is implemented using a multilayer neural network that is trained using backpropagation. An aspect of the approach is the use of an inconsistency measure to eliminate incorrect data and estimate components of the pose vector. The experimental data reported in this paper suggests that the accuracy and flexibility of the technique is good, while the online computational cost is very low.

Details
v2026.09.13