Arrow Research search

Author name cluster

Ang Li

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.

52 papers
2 author rows

Possible papers

52

AAAI Conference 2026 Conference Paper

CoordAR: One-Reference 6D Pose Estimation of Novel Objects via Autoregressive Coordinate Map Generation

  • Dexin Zuo
  • Ang Li
  • Wei Wang
  • Wenxian Yu
  • Danping Zou

Object 6D pose estimation, a crucial task for robotics and augmented reality applications, becomes particularly challenging when dealing with novel objects whose 3D models are not readily available. To reduce dependency on 3D models, recent studies have explored one-reference-based pose estimation, which requires only a single reference view instead of a complete 3D model. However, existing methods that rely on real-valued coordinate regression suffer from limited global consistency due to the local nature of convolutional architectures and face challenges in symmetric or occluded scenarios owing to a lack of uncertainty modeling. We present CoordAR, a novel autoregressive framework for one-reference 6D pose estimation of unseen objects. CoordAR formulates 3D-3D correspondences between the reference and query views as a map of discrete tokens, which is obtained in an autoregressive and probabilistic manner. To enable accurate correspondence regression, CoordAR introduces 1) a novel coordinate map tokenization that enables probabilistic prediction over discretized 3D space; 2) a modality-decoupled encoding strategy that separately encodes RGB appearance and coordinate cues; and 3) an autoregressive transformer decoder conditioned on both position-aligned query features and the partially generated token sequence. With these novel mechanisms, CoordAR significantly outperforms existing methods on multiple benchmarks and demonstrates strong robustness to symmetry, occlusion, and other challenges in real-world tests.

PDF Details DOI

JBHI Journal 2026 Journal Article

Enhancing the Reliability of Affective Brain-Computer Interfaces by Using Specifically Designed Confidence Estimator

  • Jiaheng Wang
  • Zhenyu Wang
  • Tianheng Xu
  • Ang Li
  • Yuan Si
  • Ting Zhou
  • Xi Zhao
  • Honglin Hu

In recent years, the diverse applications of electroencephalography (EEG) - based affective brain-computer interfaces (aBCIs) are being extensively explored. However, due to adverse factors like noise and physiological variability, the recognition capability of aBCIs can unforeseeably suffer abrupt declines. Since the timing of these aBCI failures is unknown, placing trust in aBCIs without scrutiny can lead to undesirable consequences. To alleviate this issue, we propose an algorithm for estimating the reliability of aBCI (primarily Graph Convolutional Network), synchronously delivering a probabilistic confidence score upon aBCI decision completion, thereby reflecting the aBCI’s real-time recognition capabilities. Methodologically, we use the Maximum Softmax Probability (MSP) from EEG recognition networks as confidence scores and leverage the Scaling Operator to calibrate them. Then, the Projection Operator is employed to address confidence estimation biases caused by noise and subject variability. For the numerical concentration of MSP, we provide fresh insights into its causes and propose corresponding solutions. The derivation of the estimator from the Maximum Entropy Principle is also substantiated for robust theoretical underpinnings. Finally, we confirm theoretically that the estimator does not compromise BCI performance. In experiments conducted on public datasets SEED and SEED-IV, the proposed algorithm demonstrates superior performance in estimating aBCIs reliability compared to other benchmarks, and commendable adaptability to new subjects. This research has the potential to lead to more trustworthy aBCIs and advance their broader application in complex real-world scenarios.

Details DOI

AAAI Conference 2026 Conference Paper

MUSE: Multimodal Uncertainty-Based Self-Driven Evolution for Robust Physiological-Signal–Based Driver Fatigue Detection

  • Jiaheng Wang
  • Yuan Si
  • Ang Li
  • Zhenyu Wang
  • Tianheng Xu
  • Honglin Hu

Precise detection of driver mental fatigue is critical for reducing traffic accidents and enhancing road safety. Compared with vision-based detection—which is susceptible to illumination and occlusion—multimodal physiological‑signal-based approaches integrate complementary information from diverse biosignals, delivering more faithful and objective fatigue assessments. However, adverse factors such as motion artifacts and environmental noise induce ceaseless deterioration to physiological signals, which markedly degrade the performance of existing multimodal fusion methods. To address this challenge, we propose Multimodal Uncertainty-based Self-driven Evolution, MUSE, reallocating modality contributions in real time via overall uncertainty minimization, thereby enabling efficient collaborative fusion of multi‐source predictions. Theoretically, MUSE guarantees a provably bounded cumulative error, and its generalization error approaches the Bayesian‑optimal fusion as iterations progress. Operating in a closed loop without labels or manual recalibration, MUSE presents superior suitability for real‑world driving scenarios compared to supervised algorithms. On the large‑scale driving fatigue dataset SEED‑VIG, MUSE outperforms existing models in both classification and regression tasks, substantiating its robustness and practicality as a promising driving fatigue detection solution.

PDF Details DOI

TMLR Journal 2026 Journal Article

Retrospective Feature Estimation for Continual Learning

  • Nghia D. Nguyen
  • Hieu Trung Nguyen
  • Ang Li
  • Hoang Pham
  • Viet Anh Nguyen
  • Khoa D Doan

The intrinsic capability to continuously learn a changing data stream is a desideratum of deep neural networks (DNNs). However, current DNNs suffer from catastrophic forgetting, which interferes with remembering past knowledge. To mitigate this issue, existing Continual Learning (CL) approaches often retain exemplars for replay, regularize learning, or allocate dedicated capacity for new tasks. This paper investigates an unexplored direction for CL called Retrospective Feature Estimation (RFE). RFE learns to reverse feature changes by aligning the features from the current trained DNN backward to the feature space of the old task, where performing predictions is easier. This retrospective process utilizes a chain of small feature mapping networks called retrospector modules. Empirical experiments on several CL benchmarks, including CIFAR10, CIFAR100, and Tiny ImageNet, demonstrate the effectiveness and potential of this novel CL direction compared to existing representative CL methods, motivating further research into retrospective mechanisms as a principled alternative for mitigating catastrophic forgetting in CL. Code is available at: https://github.com/mail-research/retrospective-feature-estimation.

PDF Details

AAAI Conference 2026 Conference Paper

Think Then Rewrite: Reasoning Enhanced Query Rewriting for Domain Specific Retrieval

  • Ang Li
  • Yufei Shi
  • Yuxuan Si
  • Yiquan Wu
  • Ming Cai
  • Xu Tan
  • Yi Wang
  • Changlong Sun

Query rewriting is a crucial task for improving retrieval, especially in professional domains such as law and medicine, where user queries are often underspecified and ambiguous. While large language models (LLMs) offer strong understanding and generation capabilities, existing LLM-based approaches reduce the task to text transformation or expansion, neglecting reasoning to disambiguate queries, which fails to bridge the cognitive gap between user queries and specialized documents. In this paper, we propose Think-Then-Rewrite (TTR), a reinforcement learning based framework that unleashes LLMs' reasoning ability for domain-specific query rewriting. TTR introduces a contrastive mutual information reward to encourage the LLM to generate reasoning processes that effectively distinguish confusing distractors. To boost early-stage training, TTR also constructs golden query rewrites as off‑policy data, providing strong guidance for RL learning. A mixed-policy optimization then combines on-policy and off-policy signals, ensuring both effectiveness and stability. Extensive experiments on legal and medical retrieval benchmarks demonstrate that TTR achieves state-of-the-art performance.

PDF Details DOI

TMLR Journal 2026 Journal Article

Uncovering the Redundancy in Transformers via a Unified Study of Layer Dropping

  • Shwai He
  • Guoheng Sun
  • Zheyu Shen
  • Ang Li

While scaling Transformer-based large language models (LLMs) has demonstrated promising performance across various tasks, it also introduces redundant architectures, posing efficiency challenges for real-world deployment. Despite some recognition of redundancy in LLMs, the variability of redundancy across different architectures in transformers, such as MLP and Attention layers, is under-explored. In this work, we investigate redundancy across different Transformer modules, including blocks, MLP layers, and attention layers, through the lens of layer dropping. Surprisingly, despite the pivotal role of attention mechanisms in distinguishing Transformers from other architectures, we find that a large portion of attention layers exhibit excessively high redundancy and can be pruned without degrading performance. For example, LLaMA-3-70B achieves a 43.4\% speedup with only a 1.8\% drop in performance by pruning half of its attention layers. In contrast, dropping MLP layers severely impairs the model's ability to distinguish between tokens, leading to catastrophic performance degradation. Moreover, our analysis reveals that attention layer redundancy persists not only throughout training but is also evident in randomly initialized models. We attribute this redundancy to three key factors that constrain representational updates from attention layers: sparse attention patterns, over-smoothed token embeddings, and the low representational magnitude of attention outputs. Overall, our findings offer valuable insights into the internal redundancy of Transformer architectures and provide practical guidance for designing more efficient LLMs. The code is released at: https://github.com/CASE-Lab-UMD/LLM-Drop.

PDF Details

NeurIPS Conference 2025 Conference Paper

$\texttt{G1}$: Teaching LLMs to Reason on Graphs with Reinforcement Learning

  • Xiaojun Guo
  • Ang Li
  • Yifei Wang
  • Stefanie Jegelka
  • Yisen Wang

Although Large Language Models (LLMs) have demonstrated remarkable progress, their proficiency in graph-related tasks remains notably limited, hindering the development of truly general-purpose models. Previous attempts, including pretraining graph foundation models or employing supervised fine-tuning, often face challenges such as the scarcity of large-scale, universally represented graph data. We introduce $\texttt{G1}$, a simple yet effective approach demonstrating that Reinforcement Learning (RL) on synthetic graph-theoretic tasks can significantly scale LLMs' graph reasoning abilities. To enable RL training, we curate \erdos, the largest graph reasoning dataset to date comprising 50 diverse graph-theoretic tasks of varying difficulty levels, 100k training data and 5k test data, all drived from real-world graphs. With RL on \erdos, $\texttt{G1}$ obtains substantial improvements in graph reasoning, where our finetuned 3B model even outperforms Qwen2. 5-72B-Instruct (24x size). RL-trained models also show strong zero-shot generalization to unseen tasks, domains, and graph encoding schemes, including other graph-theoretic benchmarks as well as real-world node classification and link prediction tasks, without compromising general reasoning abilities. Our findings offer an efficient, scalable path for building strong graph reasoners by finetuning LLMs with RL on graph-theoretic tasks, which combines the strengths of pretrained LLM capabilities with abundant, automatically generated synthetic data, suggesting that LLMs possess graph understanding abilities that RL can elicit successfully. Our implementation is open-sourced at https: //github. com/PKU-ML/G1, with models and datasets hosted on Hugging Face collections https: //huggingface. co/collections/PKU-ML/g1-683d659e992794fc99618cf2 for broader accessibility.

PDF Details

ICLR Conference 2025 Conference Paper

Agent S: An Open Agentic Framework that Uses Computers Like a Human

  • Saaket Agashe
  • Jiuzhou Han
  • Shuyu Gan
  • Jiachen Yang
  • Ang Li
  • Xin Eric Wang

We present Agent S, an open agentic framework that enables autonomous interaction with computers through Graphical User Interface (GUI), aimed at transforming human-computer interaction by automating complex, multi-step tasks. Agent S addresses three key challenges in automating computer tasks: acquiring domain-specific knowledge, planning over long task horizons, and handling dynamic, non-uniform interfaces. To this end, Agent S introduces experience-augmented hierarchical planning, which learns from external knowledge search and internal experience retrieval at multiple levels, facilitating efficient task planning and subtask execution. In addition, it employs an Agent-Computer Interface (ACI) to better elicit the reasoning and control capabilities of GUI agents based on Multimodal Large Language Models (MLLMs). Evaluation on the OSWorld benchmark shows that Agent S outperforms the baseline by 9.37\% on success rate (an 83.6\% relative improvement) and achieves a new state-of-the-art. Comprehensive analysis highlights the effectiveness of individual components and provides insights for future improvements. Furthermore, Agent S demonstrates broad generalizability to different operating systems on a newly-released WindowsAgentArena benchmark. Code available at https://github.com/simular-ai/Agent-S.

Details

JBHI Journal 2025 Journal Article

An Online Adaptation Framework for Enhancing Calibration-Free SSVEP-Based BCI Performance

  • Weize Chen
  • Jie Mei
  • Xiaolin Xiao
  • Ang Li
  • Lingling Tao
  • Kun Wang
  • Minpeng Xu
  • Dong Ming

Accomplishing a plug-and-play steady-state visual evoked potential (SSVEP)-based brain-computer interface (BCI) remains a critical challenge, due to the unsatisfying performance of calibration-free decoding algorithms. A current method called online adaptive canonical correlation analysis (OACCA) has proved efficient in enhancing calibration-free performance by self-adaptation merely with online data. However, OACCA only concerns the adaptation of spatial filters and excludes other useful adaptive procedures like individual template estimation, hindering fully exploitable model decoding and adaptation. This study proposes a new online adaptation framework termed online adaptive extended correlation analysis (OAECA) to augment the calibration-free online adaptation loop. OAECA first recalls and cleans the online trials for reliable data learning, then tunes individual templates and spatial filters for complete model updating, and finally adopts extended feature matching to improve target recognition. The simulation results on two public SSVEP datasets revealed that OAECA significantly outperformed OACCA for almost all 105 subjects, and both offline and online experiments further confirmed the effectiveness of OAECA. Particularly, OAECA achieved the highest average information transfer rate (ITR) of 202. 17 bits/min in the online experiment, significantly exceeding the state-of-the-art OACCA of 177. 02 bits/min. This study enhanced the calibration-free performance through comprehensive online adaptation, hopefully advancing SSVEP-based BCIs toward practical plug-and-play real-world applications.

Details DOI

EAAI Journal 2025 Journal Article

An undersampling method for software defect prediction based on Hilbert curve mapping distance

  • Yu Tang
  • Ye Du
  • Ang Li
  • Ming-song Yang
  • Yan Xia

The class imbalance problem presents a significant challenge in software defect prediction. The undersampling method enhances prediction performance by eliminating non-defective instances, thereby enabling the model to focus more on defective instances. However, the effective selection of representative non-defective instances while preserving the overall data distribution remains a critical challenge. Inspired by the space-filling property of Hilbert curves, we propose the Hilbert Curve Mapping Distance Undersampling (HCMDU) method for software defect prediction. This method first maps instances to Hamming space to ensure that similar instances are positioned closer together in the space. Instance circular domains are then partitioned based on the Hamming distance between them, which facilitates the exploration of instance variability within a localized region. Finally, the Hilbert curve mapping distance is employed to further uncover the data distribution pattern within the instance circular domains. The experimental results demonstrate that HCMDU delivers outstanding performance across 16 randomly selected software defect datasets in both Random Forest (RF) and Classification and Regression Trees (CART). Moreover, the results are further corroborated by the Friedman ranking and Nemenyi post-hoc test, which indicate that HCMDU significantly improves the performance of software defect prediction.

Details DOI

JBHI Journal 2025 Journal Article

BSAN: A Self-Adapted Motor Imagery Decoding Framework Based on Contextual Information

  • Zikai Wang
  • Ang Li
  • Zhenyu Wang
  • Ting Zhou
  • Tianheng Xu
  • Honglin Hu

In motor imagery (MI) decoding, it still remains challenging to excavate enough contextual information of MI in different brain regions and to bridge the cross-session variance in feature distributions. In light of these issues, our study presents an innovative Bi-Stream Adaptation Network (BSAN) to bolster network efficacy, aiming to improve MI-based brain-computer interface (BCI) robustness across sessions. Our framework consists of the Bi-attention module, feature extractor, classifier, and Bi-discriminator. Precisely, we devise the Bi-attention module to reveal granular context information of MI with performing multi-scale convolutions asymptotically. Then, after features extraction, Bi-discriminator is involved to align the features from different MI sessions such that a uniform and accurate representation of neural patterns is achieved. By such a workflow, the proposed BSAN allows for the effective fusion of context coherence and session-invariance within the network architecture, therefore diminishing the reliance of redundant MI trials for MI-BCI re-calibration. To empirically substantiate BSAN, comprehensive experiments are conducted based on two public MI datasets. With average accuracies of 78. 97% and 83. 79% on two public datasets, and an inference time of 2. 99 ms on CPU-only devices, it is believed that our approach has the potential to accelerate the practical deployment of MI-BCI.

Details DOI

NeurIPS Conference 2025 Conference Paper

FAIR Universe HiggsML Uncertainty Dataset and Competition

  • Wahid Bhimji
  • Ragansu Chakkappai
  • Po-Wen Chang
  • Yuan-Tang Chou
  • Sascha Diefenbacher
  • Jordan Dudley
  • Ibrahim Elsharkawy
  • Steven Farrell

The FAIR Universe – HiggsML Uncertainty Challenge focused on measuring the physical properties of elementary particles with imperfect simulators. Participants were required to compute and report confidence intervals for a parameter of interest regarding the Higgs boson while accounting for various systematic (epistemic) uncertainties. The dataset is a tabular dataset of 28 features and 280 million instances. Each instance represents a simulated proton-proton collision as observed at CERN’s Large Hadron Collider in Geneva, Switzerland. The features of these simulations were chosen to capture key characteristics of different types of particles. These include primary attributes, such as the energy and three-dimensional momentum of the particles, as well as derived attributes, which are calculated from the primary ones using domain-specific knowledge. Additionally, a label feature designates each instance’s type of proton-proton collision, distinguishing the Higgs boson events of interest from three background sources. As outlined in this paper, the permanent dataset release allows long-term benchmarking of new techniques. The leading submissions, including Contrastive Normalising Flows and Density Ratios estimation through classification, are described. Our challenge has brought together the physics and machine learning communities to advance our understanding and methodologies in handling systematic uncertainties within AI techniques.

PDF Details

ICLR Conference 2025 Conference Paper

PersonalLLM: Tailoring LLMs to Individual Preferences

  • Thomas P. Zollo
  • Andrew Wei Tung Siah
  • Naimeng Ye
  • Ang Li
  • Hongseok Namkoong

As LLMs become capable of complex tasks, there is growing potential for personalized interactions tailored to the subtle and idiosyncratic preferences of the user. We present a public benchmark, PersonalLLM, focusing on adapting LLMs to provide maximal benefits for a particular user. Departing from existing alignment benchmarks that implicitly assume uniform preferences, we curate open-ended prompts paired with many high-quality answers over which users would be expected to display heterogeneous latent preferences. Instead of persona prompting LLMs based on high-level attributes (e.g., user race or response length), which yields homogeneous preferences relative to humans, we develop a method that can simulate a large user base with diverse preferences from a set of pre-trained reward models. Our dataset and generated personalities offer an innovative testbed for developing personalization algorithms that grapple with continual data sparsity---few relevant feedback from the particular user---by leveraging historical data from other (similar) users. We explore basic in-context learning and meta-learning baselines to illustrate the utility of PersonalLLM and highlight the need for future methodological development.

Details

AAAI Conference 2025 Conference Paper

Reinforcement Active Client Selection for Federated Heterogeneous Graph Learning

  • Jia Wang
  • Yawen Li
  • Yingxia Shao
  • Zhe Xue
  • Zeli Guan
  • Ang Li
  • Guanhua Ye

Carefully selecting clients to participate in aggregation can assist the global model in achieving better performance. However, existing research on federated heterogeneous graph learning (FHGL) has shown limited attention to the client selection (CS) problem. Current CS algorithms face challenges in accurately evaluating client contributions and selecting appropriate participants in the context of FHGL, leading to a dilemma between convergence and accuracy. In this paper, we propose a Reinforcement Active client selection based Federated Heterogeneous Graph Learning (RAFHGL), which precisely evaluates the importance of local heterogeneous graph data and selects high-contributing clients for aggregation. RAFHGL employs an active learning agent to select representative nodes for local training. The statistical features of the active scores are used to assess client contributions. A client selection agent then chooses clients conducive to global model convergence for aggregation. To address heterogeneity introduced by sample and client selection, the training process stabilizes by correcting local losses based on data prototypes. Experimental results on 4 publicly available heterogeneous graph datasets show that RAFHGL outperforms existing Client Selection algorithms in federated heterogeneous graph learning scenarios in terms of performance and convergence.

PDF Details DOI

ICML Conference 2025 Conference Paper

Speculate, then Collaborate: Fusing Knowledge of Language Models during Decoding

  • Ziyao Wang
  • Muneeza Azmat
  • Ang Li
  • Raya Horesh
  • Mikhail Yurochkin

Large Language Models (LLMs) often excel in specific domains but fall short in others due to the limitations of their training. Thus, enabling LLMs to solve problems collaboratively by integrating their complementary knowledge promises to improve their performance across domains. To realize this potential, we introduce a novel Collaborative Speculative Decoding (CoSD) algorithm that enables efficient LLM knowledge fusion at test time without requiring additional model training. CoSD employs a draft model to generate initial sequences and an easy-to-learn rule or decision tree to decide when to invoke an assistant model to improve these drafts. CoSD not only enhances knowledge fusion but also improves inference efficiency, is transferable across domains, and offers greater explainability. Experimental results demonstrate that CoSD improves accuracy by up to 10% across benchmarks compared to existing methods, providing a scalable and effective solution for LLM-based applications. Our code has been released at https: //github. com/ATP-1010/CoSD.

Details

NeurIPS Conference 2025 Conference Paper

SymRTLO: Enhancing RTL Code Optimization with LLMs and Neuron-Inspired Symbolic Reasoning

  • Yiting Wang
  • Wanghao Ye
  • Ping Guo
  • Yexiao He
  • Ziyao Wang
  • Bowei Tian
  • Shwai He
  • Guoheng Sun

Optimizing Register Transfer Level (RTL) code is crucial for improving the efficiency and performance of digital circuits in the early stages of synthesis. Manual rewriting, guided by synthesis feedback, can yield high-quality results but is time-consuming and error-prone. Most existing compiler-based approaches have difficulty handling complex design constraints. Large Language Model (LLM)-based methods have emerged as a promising alternative to address these challenges. However, LLM-based approaches often face difficulties in ensuring alignment between the generated code and the provided prompts. This paper introduces SymRTLO, a neuron-symbolic framework that integrates LLMs with symbolic reasoning for the efficient and effective optimization of RTL code. Our method incorporates a retrieval-augmented system of optimization rules and Abstract Syntax Tree (AST)-based templates, enabling LLM-based rewriting that maintains syntactic correctness while minimizing undesired circuit behaviors. A symbolic module is proposed for analyzing and optimizing finite state machine (FSM) logic, allowing fine-grained state merging and partial specification handling beyond the scope of pattern-based compilers. Furthermore, a fast verification pipeline, combining formal equivalence checks with test-driven validation, further reduces the complexity of verification. Experiments on the RTL-Rewriter benchmark with Synopsys Design Compiler and Yosys show that SymRTLO improves power, performance, and area (PPA) by up to 43. 9%, 62. 5%, and 51. 1%, respectively, compared to the state-of-the-art methods. We will release the code as open source upon the paper's acceptance.

PDF Details

TMLR Journal 2025 Journal Article

Towards Efficient Mixture of Experts: A Holistic Study of Compression Techniques

  • Shwai He
  • Daize Dong
  • Liang Ding
  • Ang Li

Scaling large language models has driven remarkable advancements across various domains, yet the continual increase in model size presents significant challenges for real-world deployment. The Mixture of Experts (MoE) architecture offers a promising solution by dynamically selecting and activating only a subset of experts during inference, thus substantially reducing computational costs while preserving high performance. Despite these benefits, MoE introduces new inefficiencies, such as excessive parameters and communication overhead. In this work, we present a holistic study of compression techniques for Mixture of Experts to enhance both efficiency and scalability. While recent efforts have focused on Expert Trimming, which reduces the number of experts, these approaches still suffer from considerable communication and computational costs. To address this, we propose more aggressive strategies, such as Layer Drop, which removes entire MoE layers, and Block Drop, which eliminates transformer blocks. Surprisingly, these aggressive pruning techniques not only preserve model performance but also substantially improve computation and memory efficiency. Furthermore, beyond Expert Trimming, we also introduce Expert Slimming, which compresses individual experts to further boost performance and can be seamlessly integrated with Expert Trimming. Extensive experimental results demonstrate the effectiveness of our proposed methods—Layer Drop and Block Drop—along with the comprehensive recipe that integrates Expert Slimming and Expert Trimming, achieving a 6.05× speedup with 77.1% reduced memory usage while maintaining over 92% of performance on Mixtral-8×7B. Our code is released at https://github.com/CASE-Lab-UMD/Unified-MoE-Compression.

PDF Details

NeurIPS Conference 2024 Conference Paper

$\texttt{Model-GLUE}$: Democratized LLM Scaling for A Large Model Zoo in the Wild

  • Xinyu Zhao
  • Guoheng Sun
  • Ruisi Cai
  • Yukun Zhou
  • Pingzhi Li
  • Peihao Wang
  • Bowen Tan
  • Yexiao He

As Large Language Models (LLMs) excel across tasks and specialized domains, scaling LLMs based on existing models has gained significant attention, which is challenged by potential performance drop when combining disparate models. Various techniques have been proposed to aggregate pre-trained LLMs, including model merging, Mixture-of-Experts, and stacking. Despite their merits, a comprehensive comparison and synergistic application of them to a diverse model zoo is yet to be adequately addressed. In light of this research gap, this paper introduces $\texttt{Model-GLUE}$, a holistic LLM scaling guideline. First, our work starts with a benchmarking of existing LLM scaling techniques, especially selective merging, and variants of mixture. Utilizing the insights from the benchmark results, we formulate a strategy for the selection and aggregation of a heterogeneous model zoo characterizing different architectures and initialization. Our methodology involves clustering mergeable models, selecting a merging strategy, and integrating model clusters through model-level mixture. Finally, evidenced by our experiments on a diverse Llama-2-based model zoo, $\texttt{Model-GLUE}$ shows an average performance enhancement of 5. 61\%, achieved without additional training. Codes are available at https: //github. com/Model-GLUE/Model-GLUE.

PDF Details DOI

ICLR Conference 2024 Conference Paper

Enhancing One-Shot Federated Learning Through Data and Ensemble Co-Boosting

  • Rong Dai
  • Yonggang Zhang 0003
  • Ang Li
  • Tongliang Liu
  • Xun Yang 0001
  • Bo Han 0003

One-shot Federated Learning (OFL) has become a promising learning paradigm, enabling the training of a global server model via a single communication round. In OFL, the server model is aggregated by distilling knowledge from all client models (the ensemble), which are also responsible for synthesizing samples for distillation. In this regard, advanced works show that the performance of the server model is intrinsically related to the quality of the synthesized data and the ensemble model. To promote OFL, we introduce a novel framework, Co-Boosting, in which synthesized data and the ensemble model mutually enhance each other progressively. Specifically, Co-Boosting leverages the current ensemble model to synthesize higher-quality samples in an adversarial attack manner. These hard samples are then employed to promote the quality of the ensemble model by adjusting the ensembling weights for each client model. Consequently, Co-Boosting periodically achieves high-quality data and ensemble models. Extensive experiments demonstrate that Co-Boosting can substantially outperform existing baselines under various settings. Moreover, Co-Boosting eliminates the need for adjustments to the client's local training, requires no additional data or model transmission, and allows client models to have heterogeneous architectures.

Details

NeurIPS Conference 2024 Conference Paper

FLoRA: Federated Fine-Tuning Large Language Models with Heterogeneous Low-Rank Adaptations

  • Ziyao Wang
  • Zheyu Shen
  • Yexiao He
  • Guoheng Sun
  • Hongyi Wang
  • Lingjuan Lyu
  • Ang Li

The rapid development of Large Language Models (LLMs) has been pivotal in advancing AI, with pre-trained LLMs being adaptable to diverse downstream tasks through fine-tuning. Federated learning (FL) further enhances fine-tuning in a privacy-aware manner by utilizing clients' local data through in-situ computation, eliminating the need for data movement. However, fine-tuning LLMs, given their massive scale of parameters, poses challenges for clients with constrained and heterogeneous resources in FL. Previous methods employed low-rank adaptation (LoRA) for efficient federated fine-tuning but utilized traditional FL aggregation strategies on LoRA adapters. This approach led to mathematically inaccurate aggregation noise, reducing fine-tuning effectiveness and failing to address heterogeneous LoRAs. In this work, we first highlight the mathematical incorrectness of LoRA aggregation in existing federated fine-tuning methods. We introduce a new approach called FLoRA that enables federated fine-tuning on heterogeneous LoRA adapters across clients through a novel stacking-based aggregation method. Our approach is noise-free and seamlessly supports heterogeneous LoRAs. Extensive experiments demonstrate FLoRA's superior performance in both homogeneous and heterogeneous settings, surpassing state-of-the-art methods. We envision this work as a milestone for efficient, privacy-preserving, and accurate federated fine-tuning of LLMs.

PDF Details DOI

EAAI Journal 2024 Journal Article

Geotechnical risk modeling using an explainable transfer learning model incorporating physical guidance

  • Fenghua Liu
  • Wenli Liu
  • Ang Li
  • Jack C.P. Cheng

While Artificial intelligence (AI) has been successfully applied in assessing geotechnical risk, such methods heavily rely on data quality to achieve satisfactory performance, and their results hardly can be interpreted due to their opaque design. With this in mind, this paper aims to address the following research gap: How can we accurately model geotechnical risks using limited data and domain knowledge, and efficiently explain the results of AI model? We develop a physics-guided transfer learning (PGTL) model to enhance the explainability and accuracy of geotechnical risk modeling. With the help of a physical model that simulates the tunnel excavation, a physics-guided dataset with 1000 samples is established and used to train a deep neural network. On these bases, transfer learning is adopted to fuse the features of physics mechanisms and monitoring data, constructing an explainable prediction model of geotechnical risk. To further support risk decision-making, feature relevance techniques are employed to assess the contribution of input parameters to risk. A shield tunnel construction in Wuhan is selected as a case to validate the effectiveness of the proposed method. The PGTL exhibits a more promising accuracy with R 2 of 0. 777 in contrast to three popular machine learning approaches, and provides insights into parameters that significantly induce risk, enhancing site managers’ understanding of tunnel construction and being conducive to tunnel safety.

Details DOI

ICRA Conference 2024 Conference Paper

Ground-Fusion: A Low-cost Ground SLAM System Robust to Corner Cases

  • Jie Yin
  • Ang Li
  • Wei Xi
  • Wenxian Yu
  • Danping Zou

We introduce Ground-Fusion, a low-cost sensor fusion simultaneous localization and mapping (SLAM) system for ground vehicles. Our system features efficient initialization, effective sensor anomaly detection and handling, real-time dense color mapping, and robust localization in diverse environments. We tightly integrate RGB-D images, inertial measurements, wheel odometer and GNSS signals within a factor graph to achieve accurate and reliable localization both indoors and outdoors. To ensure successful initialization, we propose an efficient strategy that comprises three different methods: stationary, visual, and dynamic, tailored to handle diverse cases. Furthermore, we develop mechanisms to detect sensor anomalies and degradation, handling them adeptly to maintain system accuracy. Our experimental results on both public and self-collected datasets demonstrate that Ground-Fusion outperforms existing low-cost SLAM systems in corner cases. We release the code and datasets at https://github.com/SJTU-ViSYS/Ground-Fusion.

Details

ICML Conference 2024 Conference Paper

PID: Prompt-Independent Data Protection Against Latent Diffusion Models

  • Ang Li
  • Yichuan Mo
  • Mingjie Li 0007
  • Yisen Wang 0001

The few-shot fine-tuning of Latent Diffusion Models (LDMs) has enabled them to grasp new concepts from a limited number of images. However, given the vast amount of personal images accessible online, this capability raises critical concerns about civil privacy. While several previous defense methods have been developed to prevent such misuse of LDMs, they typically assume that the textual prompts used by data protectors exactly match those employed by data exploiters. In this paper, we first empirically demonstrate that breaking this assumption, i. e. , in cases where discrepancies exist between the textual conditions used by protectors and exploiters, could substantially reduces the effectiveness of these defenses. Furthermore, considering the visual encoder’s independence from textual prompts, we delve into the visual encoder and thoroughly investigate how manipulating the visual encoder affects the few-shot fine-tuning process of LDMs. Drawing on these insights, we propose a simple yet effective method called Prompt-Independent Defense (PID) to safeguard privacy against LDMs. We show that PID can act as a strong privacy shield on its own while requiring significantly less computational power. We believe our studies, along with the comprehensive understanding and new defense method, provide a notable advance toward reliable data protection against LDMs.

Details

AAAI Conference 2024 Conference Paper

Probabilities of Causation with Nonbinary Treatment and Effect

  • Ang Li
  • Judea Pearl

Probabilities of causation are proven to be critical in modern decision-making. This paper deals with the problem of estimating the probabilities of causation when treatment and effect are not binary. Pearl defined the binary probabilities of causation, such as the probability of necessity and sufficiency (PNS), the probability of sufficiency (PS), and the probability of necessity (PN). Tian and Pearl then derived sharp bounds for these probabilities of causation using experimental and observational data. In this paper, we define and provide theoretical bounds for all types of probabilities of causation with multivalued treatments and effects. We further discuss examples where our bounds guide practical decisions and use simulation studies to evaluate how informative the bounds are for various data combinations.

PDF Details DOI

NeurIPS Conference 2024 Conference Paper

QGym: Scalable Simulation and Benchmarking of Queuing Network Controllers

  • Haozhe Chen
  • Ang Li
  • Ethan Che
  • Tianyi Peng
  • Jing Dong
  • Hongseok Namkoong

Queuing network control allows allocation of scarce resources to manage congestion, a fundamental problem in manufacturing, communications, and healthcare. Compared to standard RL problems, queueing problems are distinguished by unique challenges: i) a system operating in continuous time, ii) high stochasticity, and iii) long horizons over which the system can become unstable (exploding delays). To provide the empirical foundations for methodological development tackling these challenges, we present an open-sourced queueing simulation framework, QGym, that benchmark queueing policies across realistic problem instances. Our modular framework allows the researchers to build on our initial instances, which provide a wide range of environments including parallel servers, criss-cross, tandem, and re-entrant networks, as well as a realistically calibrated hospital queuing system. From these, various policies can be easily tested, including both model-free RL methods and classical queuing policies. Our testbed significantly expands the scope of empirical benchmarking in prior work, and complements thetraditional focus on evaluating algorithms based on mathematical guarantees in idealized settings. QGym code is open-sourced at https: //github. com/namkoong-lab/QGym.

PDF Details DOI

NeurIPS Conference 2024 Conference Paper

SHED: Shapley-Based Automated Dataset Refinement for Instruction Fine-Tuning

  • Yexiao He
  • Ziyao Wang
  • Zheyu Shen
  • Guoheng Sun
  • Yucong Dai
  • Yongkai Wu
  • Hongyi Wang
  • Ang Li

The pre-trained Large Language Models (LLMs) can be adapted for many downstream tasks and tailored to align with human preferences through fine-tuning. Recent studies have discovered that LLMs can achieve desirable performance with only a small amount of high-quality data, suggesting that a large portion of the data in these extensive datasets is redundant or even harmful. Identifying high-quality data from vast datasets to curate small yet effective datasets has emerged as a critical challenge. In this paper, we introduce SHED, an automated dataset refinement framework based on Shapley value for instruction fine-tuning. SHED eliminates the need for human intervention or the use of commercial LLMs. Moreover, the datasets curated through SHED exhibit transferability, indicating they can be reused across different LLMs with consistently high performance. We conduct extensive experiments to evaluate the datasets curated by SHED. The results demonstrate SHED's superiority over state-of-the-art methods across various tasks and LLMs; notably, datasets comprising only 10% of the original data selected by SHED achieve performance comparable to or surpassing that of the full datasets.

PDF Details DOI

SoCS Conference 2024 Conference Paper

Solving Facility Location Problems via FastMap and Locality Sensitive Hashing

  • Ang Li
  • Peter J. Stuckey
  • Sven Koenig
  • T. K. Satish Kumar

Facility Location Problems (FLPs) arise while serving multiple customers in a shared environment, minimizing transportation and other costs. Hence, they involve the optimal placement of facilities. They are defined on graphs as well as in Euclidean spaces with or without obstacles; and they are typically NP-hard to solve optimally. There are many heuristic algorithms tailored to different kinds of FLPs. However, FLPs defined in Euclidean spaces without obstacles are the most amenable to efficient and effective heuristic algorithms. This motivates the idea of quickly reformulating FLPs on graphs and in Euclidean spaces with obstacles to FLPs in Euclidean spaces without obstacles. Towards this end, we propose a new approach that uses FastMap and Locality Sensitive Hashing. FastMap is a near-linear-time algorithm that embeds the vertices of a graph in a Euclidean space while approximately preserving graph-based distances as Euclidean distances for all pairs of vertices. Through extensive experiments, we show that our approach significantly outperforms other state-of-the-art competing algorithms on a variety of FLPs: the Multi-Agent Meeting, Vertex K-Median (VKM), Weighted VKM, and the Capacitated VKM problems.

Details

ICRA Conference 2024 Conference Paper

Stereo-LiDAR Depth Estimation with Deformable Propagation and Learned Disparity-Depth Conversion

  • Ang Li
  • Anning Hu
  • Wei Xi
  • Wenxian Yu
  • Danping Zou

Accurate and dense depth estimation with stereo cameras and LiDAR is an important task for automatic driving and robotic perception. While sparse hints from LiDAR points have improved cost aggregation in stereo matching, their effectiveness is limited by the low density and non-uniform distribution. To address this issue, we propose a novel stereo-LiDAR depth estimation network with Semi-Dense hint Guidance, named SDG-Depth. Our network includes a deformable propagation module for generating a semi-dense hint map and a confidence map by propagating sparse hints using a learned deformable window. These maps then guide cost aggregation in stereo matching. To reduce the triangulation error in depth recovery from disparity, especially in distant regions, we introduce a disparity-depth conversion module. Our method is both accurate and efficient. The experimental results on benchmark tests show its superior performance. Our code is available at https://github.com/SJTU-ViSYS/SDG-Depth.

Details

ICML Conference 2024 Conference Paper

TERD: A Unified Framework for Safeguarding Diffusion Models Against Backdoors

  • Yichuan Mo
  • Hui Huang
  • Mingjie Li 0007
  • Ang Li
  • Yisen Wang 0001

Diffusion models have achieved notable success in image generation, but they remain highly vulnerable to backdoor attacks, which compromise their integrity by producing specific undesirable outputs when presented with a pre-defined trigger. In this paper, we investigate how to protect diffusion models from this dangerous threat. Specifically, we propose TERD, a backdoor defense framework that builds unified modeling for current attacks, which enables us to derive an accessible reversed loss. A trigger reversion strategy is further employed: an initial approximation of the trigger through noise sampled from a prior distribution, followed by refinement through differential multi-step samplers. Additionally, with the reversed trigger, we propose backdoor detection from the noise space, introducing the first backdoor input detection approach for diffusion models and a novel model detection algorithm that calculates the KL divergence between reversed and benign distributions. Extensive evaluations demonstrate that TERD secures a 100% True Positive Rate (TPR) and True Negative Rate (TNR) across datasets of varying resolutions. TERD also demonstrates nice adaptability to other Stochastic Differential Equation (SDE)-based models. Our code is available at https: //github. com/PKU-ML/TERD.

Details

AAAI Conference 2024 Conference Paper

Unit Selection with Nonbinary Treatment and Effect

  • Ang Li
  • Judea Pearl

The unit selection problem aims to identify a set of individuals who are most likely to exhibit a desired mode of behavior or to evaluate the percentage of such individuals in a given population, for example, selecting individuals who would respond one way if encouraged and a different way if not encouraged. Using a combination of experimental and observational data, Li and Pearl solved the binary unit selection problem (binary treatment and effect) by deriving tight bounds on the "benefit function," which is the payoff/cost associated with selecting an individual with given characteristics. This paper extends the benefit function to the general form such that the treatment and effect are not restricted to binary. We then propose an algorithm to test the identifiability of the nonbinary benefit function and an algorithm to compute the bounds of the nonbinary benefit function using experimental and observational data.

PDF Details DOI

NeurIPS Conference 2023 Conference Paper

Adversarial Examples Are Not Real Features

  • Ang Li
  • Yifei Wang
  • Yiwen Guo
  • Yisen Wang

The existence of adversarial examples has been a mystery for years and attracted much interest. A well-known theory by \citet{ilyas2019adversarial} explains adversarial vulnerability from a data perspective by showing that one can extract non-robust features from adversarial examples and these features alone are useful for classification. However, the explanation remains quite counter-intuitive since non-robust features are mostly noise features to humans. In this paper, we re-examine the theory from a larger context by incorporating multiple learning paradigms. Notably, we find that contrary to their good usefulness under supervised learning, non-robust features attain poor usefulness when transferred to other self-supervised learning paradigms, such as contrastive learning, masked image modeling, and diffusion models. It reveals that non-robust features are not really as useful as robust or natural features that enjoy good transferability between these paradigms. Meanwhile, for robustness, we also show that naturally trained encoders from robust features are largely non-robust under AutoAttack. Our cross-paradigm examination suggests that the non-robust features are not really useful but more like paradigm-wise shortcuts, and robust features alone might be insufficient to attain reliable model robustness. Code is available at \url{https: //github. com/PKU-ML/AdvNotRealFeatures}.

PDF Details

NeurIPS Conference 2023 Conference Paper

FedNAR: Federated Optimization with Normalized Annealing Regularization

  • Junbo Li
  • Ang Li
  • Chong Tian
  • Qirong Ho
  • Eric Xing
  • Hongyi Wang

Weight decay is a standard technique to improve generalization performance in modern deep neural network optimization, and is also widely adopted in federated learning (FL) to prevent overfitting in local clients. In this paper, we first explore the choices of weight decay and identify that weight decay value appreciably influences the convergence of existing FL algorithms. While preventing overfitting is crucial, weight decay can introduce a different optimization goal towards the global objective, which is further amplified in FL due to multiple local updates and heterogeneous data distribution. To address this challenge, we develop {\it Federated optimization with Normalized Annealing Regularization} (FedNAR), a simple yet effective and versatile algorithmic plug-in that can be seamlessly integrated into any existing FL algorithms. Essentially, we regulate the magnitude of each update by performing co-clipping of the gradient and weight decay. We provide a comprehensive theoretical analysis of FedNAR's convergence rate and conduct extensive experiments on both vision and language datasets with different backbone federated optimization algorithms. Our experimental results consistently demonstrate that incorporating FedNAR into existing FL algorithms leads to accelerated convergence and heightened model accuracy. Moreover, FedNAR exhibits resilience in the face of various hyperparameter configurations. Specifically, FedNAR has the ability to self-adjust the weight decay when the initial specification is not optimal, while the accuracy of traditional FL algorithms would markedly decline. Our codes are released at \href{https: //anonymous. 4open. science/r/fednar-BE8F}{https: //anonymous. 4open. science/r/fednar-BE8F}.

PDF Details

AAAI Conference 2023 Conference Paper

Ising-Traffic: Using Ising Machine Learning to Predict Traffic Congestion under Uncertainty

  • Zhenyu Pan
  • Anshujit Sharma
  • Jerry Yao-Chieh Hu
  • Zhuo Liu
  • Ang Li
  • Han Liu
  • Michael Huang
  • Tony Geng

This paper addresses the challenges in accurate and real-time traffic congestion prediction under uncertainty by proposing Ising-Traffic, a dual-model Ising-based traffic prediction framework that delivers higher accuracy and lower latency than SOTA solutions. While traditional solutions face the dilemma from the trade-off between algorithm complexity and computational efficiency, our Ising-based method breaks away from the trade-off leveraging the Ising model's strong expressivity and the Ising machine's strong computation power. In particular, Ising-Traffic formulates traffic prediction under uncertainty into two Ising models: Reconstruct-Ising and Predict-Ising. Reconstruct-Ising is mapped onto modern Ising machines and handles uncertainty in traffic accurately with negligible latency and energy consumption, while Predict-Ising is mapped onto traditional processors and predicts future congestion precisely with only at most 1.8% computational demands of existing solutions. Our evaluation shows Ising-Traffic delivers on average 98X speedups and 5% accuracy improvement over SOTA.

PDF Details DOI

ICLR Conference 2022 Conference Paper

AEVA: Black-box Backdoor Detection Using Adversarial Extreme Value Analysis

  • Junfeng Guo
  • Ang Li
  • Cong Liu 0005

Deep neural networks (DNNs) are proved to be vulnerable against backdoor attacks. A backdoor could be embedded in the target DNNs through injecting a backdoor trigger into the training examples, which can cause the target DNNs misclassify an input attached with the backdoor trigger. Recent backdoor detection methods often require the access to the original poisoned training data, the parameters of the target DNNs, or the predictive confidence for each given input, which are impractical in many real-world applications, e.g., on-device de-ployed DNNs. We address the black-box hard-label backdoor detection problem where the DNN is a fully black-box and only its final output label is accessible. We approach this problem from the optimization perspective and show that the objective of backdoor detection is bounded by an adversarial objective. Further theoretical and empirical studies reveal that this adversarial objective leads to a solution with highly skewed distribution; a singularity is often observed in the adversarial map of a backdoor-infected example, which we call the adversarial singularity phenomenon. Based on this observation, we propose the adversarial extreme value analysis(AEVA) algorithm to detect backdoors in black-box neural networks. The AEVA algorithm is based on an extreme value analysis on the adversarial map, computed from the monte-carlo gradient estimation due to the black-box hard-label constraint. Evidenced by extensive experiments across three popular tasks and backdoor attacks, our approach is shown effective in detecting backdoor attacks under the black-box hard-label scenarios

Details

AAAI Conference 2022 Conference Paper

Bounds on Causal Effects and Application to High Dimensional Data

  • Ang Li
  • Judea Pearl

This paper addresses the problem of estimating causal effects when adjustment variables in the back-door or front-door criterion are partially observed. For such scenarios, we derive bounds on the causal effects by solving two non-linear optimization problems, and demonstrate that the bounds are sufficient. Using this optimization method, we propose a framework for dimensionality reduction that allows one to trade bias for estimation power, and demonstrate its performance using simulation studies.

PDF Details

IJCAI Conference 2022 Conference Paper

Causes of Effects: Learning Individual Responses from Population Data

  • Scott Mueller
  • Ang Li
  • Judea Pearl

The problem of individualization is crucial in almost every field of science. Identifying causes of specific observed events is likewise essential for accurate decision making as well as explanation. However, such tasks invoke counterfactual relationships, and are therefore indeterminable from population data. For example, the probability of benefiting from a treatment concerns an individual having a favorable outcome if treated and an unfavorable outcome if untreated; it cannot be estimated from experimental data, even when conditioned on fine-grained features, because we cannot test both possibilities for an individual. Tian and Pearl provided bounds on this and other probabilities of causation using a combination of experimental and observational data. Those bounds, though tight, can be narrowed significantly when structural information is available in the form of a causal model. This added information may provide the power to solve central problems, such as explainable AI, legal responsibility, and personalized medicine, all of which demand counterfactual logic. This paper derives, analyzes, and characterizes these new bounds, and illustrates some of their practical applications.

PDF Details DOI

ICLR Conference 2022 Conference Paper

Information-theoretic Online Memory Selection for Continual Learning

  • Shengyang Sun
  • Daniele Calandriello
  • Huiyi Hu
  • Ang Li
  • Michalis K. Titsias

A challenging problem in task-free continual learning is the online selection of a representative replay memory from data streams. In this work, we investigate the online memory selection problem from an information-theoretic perspective. To gather the most information, we propose the \textit{surprise} and the \textit{learnability} criteria to pick informative points and to avoid outliers. We present a Bayesian model to compute the criteria efficiently by exploiting rank-one matrix structures. We demonstrate that these criteria encourage selecting informative points in a greedy algorithm for online memory selection. Furthermore, by identifying the importance of \textit{the timing to update the memory}, we introduce a stochastic information-theoretic reservoir sampler (InfoRS), which conducts sampling among selective points with high information. Compared to reservoir sampling, InfoRS demonstrates improved robustness against data imbalance. Finally, empirical performances over continual learning benchmarks manifest its efficiency and efficacy.

Details

IJCAI Conference 2022 Conference Paper

Iterative Geometry-Aware Cross Guidance Network for Stereo Image Inpainting

  • Ang Li
  • Shanshan Zhao
  • Zhang Qingjie
  • Qiuhong Ke

Currently, single image inpainting has achieved promising results based on deep convolutional neural networks. However, inpainting on stereo images with missing regions has not been explored thoroughly, which is also a significant but different problem. One crucial requirement for stereo image inpainting is stereo consistency. To achieve it, we propose an Iterative Geometry-Aware Cross Guidance Network (IGGNet). The IGGNet contains two key ingredients, i. e. , a Geometry-Aware Attention(GAA) module and an Iterative Cross Guidance(ICG) strategy. The GAA module relies on the epipolar geometry cues and learns the geometry-aware guidance from one view to another, which is beneficial to make the corresponding regions in two views consistent. However, learning guidance from co-existing missing regions is challenging. To address this issue, the ICG strategy is proposed, which can alternately narrow down the missing regions of the two views in an iterative manner. Experimental results demonstrate that our proposed network outperforms the latest stereo image inpainting model and state-of-the-art single image inpainting models.

PDF Details DOI

AAAI Conference 2022 Conference Paper

SimIPU: Simple 2D Image and 3D Point Cloud Unsupervised Pre-training for Spatial-Aware Visual Representations

  • Zhenyu Li
  • Zehui Chen
  • Ang Li
  • Liangji Fang
  • Qinhong Jiang
  • Xianming Liu
  • Junjun Jiang
  • Bolei Zhou

Pre-training has become a standard paradigm in many computer vision tasks. However, most of the methods are generally designed on the RGB image domain. Due to the discrepancy between the two-dimensional image plane and the three-dimensional space, such pre-trained models fail to perceive spatial information and serve as sub-optimal solutions for 3D-related tasks. To bridge this gap, we aim to learn a spatial-aware visual representation that can describe the three-dimensional space and is more suitable and effective for these tasks. To leverage point clouds, which are much more superior in providing spatial information compared to images, we propose a simple yet effective 2D Image and 3D Point cloud Unsupervised pre-training strategy, called SimIPU. Specifically, we develop a multi-modal contrastive learning framework that consists of an intra-modal spatial perception module to learn a spatial-aware representation from point clouds and an inter-modal feature interaction module to transfer the capability of perceiving spatial information from the point cloud encoder to the image encoder, respectively. Positive pairs for contrastive losses are established by the matching algorithm and the projection matrix. The whole framework is trained in an unsupervised end-toend fashion. To the best of our knowledge, this is the first study to explore contrastive learning pre-training strategies for outdoor multi-modal datasets, containing paired camera images and LIDAR point clouds.

PDF Details

YNIMG Journal 2022 Journal Article

The human mediodorsal thalamus: Organization, connectivity, and function

  • Kaixin Li
  • Lingzhong Fan
  • Yue Cui
  • Xuehu Wei
  • Yini He
  • Jiyue Yang
  • Yuheng Lu
  • Wen Li

The human mediodorsal thalamic nucleus (MD) is crucial for higher cognitive functions, while the fine anatomical organization of the MD and the function of each subregion remain elusive. In this study, using high-resolution data provided by the Human Connectome Project, an anatomical connectivity-based method was adopted to unveil the topographic organization of the MD. Four fine-grained subregions were identified in each hemisphere, including the medial (MDm), central (MDc), dorsal (MDd), and lateral (MDl), which recapitulated previous cytoarchitectonic boundaries from histological studies. The subsequent connectivity analysis of the subregions also demonstrated distinct anatomical and functional connectivity patterns, especially with the prefrontal cortex. To further evaluate the function of MD subregions, partial least squares analysis was performed to examine the relationship between different prefrontal-subregion connectivity and behavioral measures in 1012 subjects. The results showed subregion-specific involvement in a range of cognitive functions. Specifically, the MDm predominantly subserved emotional-cognition domains, while the MDl was involved in multiple cognitive functions especially cognitive flexibility and inhibition. The MDc and MDd were correlated with fluid intelligence, processing speed, and emotional cognition. In conclusion, our work provides new insights into the anatomical and functional organization of the MD and highlights the various roles of the prefrontal-thalamic circuitry in human cognition.

Details DOI

AAAI Conference 2022 Conference Paper

Unit Selection with Causal Diagram

  • Ang Li
  • Judea Pearl

The unit selection problem aims to identify a set of individuals who are most likely to exhibit a desired mode of behavior, for example, selecting individuals who would respond one way if encouraged and a different way if not encouraged. Using a combination of experimental and observational data, Li and Pearl derived tight bounds on the “benefit function” - the payoff/cost associated with selecting an individual with given characteristics. This paper shows that these bounds can be narrowed significantly (enough to change decisions) when structural information is available in the form of a causal model. We address the problem of estimating the benefit function using observational and experimental data when specific graphical criteria are assumed to hold.

PDF Details

NeurIPS Conference 2021 Conference Paper

FL-WBC: Enhancing Robustness against Model Poisoning Attacks in Federated Learning from a Client Perspective

  • Jingwei Sun
  • Ang Li
  • Louis DiValentin
  • Amin Hassanzadeh
  • Yiran Chen
  • Hai Li

Federated learning (FL) is a popular distributed learning framework that trains a global model through iterative communications between a central server and edge devices. Recent works have demonstrated that FL is vulnerable to model poisoning attacks. Several server-based defense approaches (e. g. robust aggregation), have been proposed to mitigate such attacks. However, we empirically show that under extremely strong attacks, these defensive methods fail to guarantee the robustness of FL. More importantly, we observe that as long as the global model is polluted, the impact of attacks on the global model will remain in subsequent rounds even if there are no subsequent attacks. In this work, we propose a client-based defense, named White Blood Cell for Federated Learning (FL-WBC), which can mitigate model poisoning attacks that have already polluted the global model. The key idea of FL-WBC is to identify the parameter space where long-lasting attack effect on parameters resides and perturb that space during local training. Furthermore, we derive a certified robustness guarantee against model poisoning attacks and a convergence guarantee to FedAvg after applying our FL-WBC. We conduct experiments on FasionMNIST and CIFAR10 to evaluate the defense against state-of-the-art model poisoning attacks. The results demonstrate that our method can effectively mitigate model poisoning attack impact on the global model within 5 communication rounds with nearly no accuracy drop under both IID and Non-IID settings. Our defense is also complementary to existing server-based robust aggregation approaches and can further improve the robustness of FL under extremely strong attacks.

PDF Details

IJCAI Conference 2021 Conference Paper

Noise Doesn't Lie: Towards Universal Detection of Deep Inpainting

  • Ang Li
  • Qiuhong Ke
  • Xingjun Ma
  • Haiqin Weng
  • Zhiyuan Zong
  • Feng Xue
  • Rui Zhang

Deep image inpainting aims to restore damaged or missing regions in an image with realistic contents. While having a wide range of applications such as object removal and image recovery, deep inpainting techniques also have the risk of being manipulated for image forgery. A promising countermeasure against such forgeries is deep inpainting detection, which aims to locate the inpainted regions in an image. In this paper, we make the first attempt towards universal detection of deep inpainting, where the detection network can generalize well when detecting different deep inpainting methods. To this end, we first propose a novel data generation approach to generate a universal training dataset, which imitates the noise discrepancies exist in real versus inpainted image contents to train universal detectors. We then design a Noise-Image Cross-fusion Network (NIX-Net) to effectively exploit the discriminative information contained in both the images and their noise patterns. We empirically show, on multiple benchmark datasets, that our approach outperforms existing detection methods by a large margin and generalize well to unseen deep inpainting techniques. Our universal training dataset can also significantly boost the generalizability of existing detection methods.

PDF Details DOI

AAAI Conference 2020 Conference Paper

Improved Knowledge Distillation via Teacher Assistant

  • Seyed Iman Mirzadeh
  • Mehrdad Farajtabar
  • Ang Li
  • Nir Levine
  • Akihiro Matsukawa
  • Hassan Ghasemzadeh

Despite the fact that deep neural networks are powerful models and achieve appealing results on many tasks, they are too large to be deployed on edge devices like smartphones or embedded sensor nodes. There have been efforts to compress these networks, and a popular method is knowledge distillation, where a large (teacher) pre-trained network is used to train a smaller (student) network. However, in this paper, we show that the student network performance degrades when the gap between student and teacher is large. Given a fixed student network, one cannot employ an arbitrarily large teacher, or in other words, a teacher can effectively transfer its knowledge to students up to a certain size, not smaller. To alleviate this shortcoming, we introduce multi-step knowledge distillation, which employs an intermediate-sized network (teacher assistant) to bridge the gap between the student and the teacher. Moreover, we study the effect of teacher assistant size and extend the framework to multi-step distillation. Theoretical analysis and extensive experiments on CIFAR-10, 100 and ImageNet datasets and on CNN and ResNet architectures substantiate the effectiveness of our proposed approach.

PDF Details

NeurIPS Conference 2020 Conference Paper

Learning to Incentivize Other Learning Agents

  • Jiachen Yang
  • Ang Li
  • Mehrdad Farajtabar
  • Peter Sunehag
  • Edward Hughes
  • Hongyuan Zha

The challenge of developing powerful and general Reinforcement Learning (RL) agents has received increasing attention in recent years. Much of this effort has focused on the single-agent setting, in which an agent maximizes a predefined extrinsic reward function. However, a long-term question inevitably arises: how will such independent agents cooperate when they are continually learning and acting in a shared multi-agent environment? Observing that humans often provide incentives to influence others' behavior, we propose to equip each RL agent in a multi-agent environment with the ability to give rewards directly to other agents, using a learned incentive function. Each agent learns its own incentive function by explicitly accounting for its impact on the learning of recipients and, through them, the impact on its own extrinsic objective. We demonstrate in experiments that such agents significantly outperform standard RL and opponent-shaping agents in challenging general-sum Markov games, often by finding a near-optimal division of labor. Our work points toward more opportunities and challenges along the path to ensure the common good in a multi-agent future.

PDF Details

ICLR Conference 2020 Conference Paper

Prediction, Consistency, Curvature: Representation Learning for Locally-Linear Control

  • Nir Levine
  • Yinlam Chow
  • Rui Shu
  • Ang Li
  • Mohammad Ghavamzadeh
  • Hung Bui

Many real-world sequential decision-making problems can be formulated as optimal control with high-dimensional observations and unknown dynamics. A promising approach is to embed the high-dimensional observations into a lower-dimensional latent representation space, estimate the latent dynamics model, then utilize this model for control in the latent space. An important open question is how to learn a representation that is amenable to existing control algorithms? In this paper, we focus on learning representations for locally-linear control algorithms, such as iterative LQR (iLQR). By formulating and analyzing the representation learning problem from an optimal control perspective, we establish three underlying principles that the learned representation should comprise: 1) accurate prediction in the observation space, 2) consistency between latent and observation space dynamics, and 3) low curvature in the latent space transitions. These principles naturally correspond to a loss function that consists of three terms: prediction, consistency, and curvature (PCC). Crucially, to make PCC tractable, we derive an amortized variational bound for the PCC loss function. Extensive experiments on benchmark domains demonstrate that the new variational-PCC learning algorithm benefits from significantly more stable and reproducible training, and leads to superior control performance. Further ablation studies give support to the importance of all three PCC components for learning a good latent space for control.

Details

IJCAI Conference 2019 Conference Paper

Generative Image Inpainting with Submanifold Alignment

  • Ang Li
  • Jianzhong Qi
  • Rui Zhang
  • Xingjun Ma
  • Kotagiri Ramamohanarao

Image inpainting aims at restoring missing regions of corrupted images, which has many applications such as image restoration and object removal. However, current GAN-based generative inpainting models do not explicitly exploit the structural or textural consistency between restored contents and their surrounding contexts. To address this limitation, we propose to enforce the alignment (or closeness) between the local data submanifolds (subspaces) around restored images and those around the original (uncorrupted) images during the learning process of GAN-based inpainting models. We exploit Local Intrinsic Dimensionality (LID) to measure, in deep feature space, the alignment between data submanifolds learned by a GAN model and those of the original data, from a perspective of both images (denoted as iLID) and local patches (denoted as pLID) of images. We then apply iLID and pLID as regularizations for GAN-based inpainting models to encourage two different levels of submanifold alignments: 1) an image-level alignment to improve structural consistency, and 2) a patch-level alignment to improve textural details. Experimental results on four benchmark datasets show that our proposed model can generate more accurate results than state-of-the-art models.

PDF Details

AAAI Conference 2019 Conference Paper

SNR: Sub-Network Routing for Flexible Parameter Sharing in Multi-Task Learning

  • Jiaqi Ma
  • Zhe Zhao
  • Jilin Chen
  • Ang Li
  • Lichan Hong
  • Ed H. Chi

Machine learning applications, such as object detection and content recommendation, often require training a single model to predict multiple targets at the same time. Multi-task learning through neural networks became popular recently, because it not only helps improve the accuracy of many prediction tasks when they are related, but also saves computation cost by sharing model architectures and low-level representations. The latter is critical for real-time large-scale machine learning systems. However, classic multi-task neural networks may degenerate significantly in accuracy when tasks are less related. Previous works (Misra et al. 2016; Yang and Hospedales 2016; Ma et al. 2018) showed that having more flexible architectures in multi-task models, either manually-tuned or softparameter-sharing structures like gating networks, helps improve the prediction accuracy. However, manual tuning is not scalable, and the previous soft-parameter sharing models are either not flexible enough or computationally expensive. In this work, we propose a novel framework called Sub- Network Routing (SNR) to achieve more flexible parameter sharing while maintaining the computational advantage of the classic multi-task neural-network model. SNR modularizes the shared low-level hidden layers into multiple layers of subnetworks, and controls the connection of sub-networks with learnable latent variables to achieve flexible parameter sharing. We demonstrate the effectiveness of our approach on a large-scale dataset YouTube8M. We show that the proposed method improves the accuracy of multi-task models while maintaining their computation efficiency.

PDF Details

IJCAI Conference 2019 Conference Paper

Unit Selection Based on Counterfactual Logic

  • Ang Li
  • Judea Pearl

The unit selection problem aims to identify a set of individuals who are most likely to exhibit a desired mode of behavior, which is defined in counterfactual terms. A typical example is that of selecting individuals who would respond one way if encouraged and a different way if not encouraged. Unlike previous works on this problem, which rely on ad-hoc heuristics, we approach this problem formally, using counterfactual logic, to properly capture the nature of the desired behavior. This formalism enables us to derive an informative selection criterion which integrates experimental and observational data. We demonstrate the superiority of this criterion over A/B-test-based approaches.

PDF Details

YNICL Journal 2017 Journal Article

Polygenic risk for five psychiatric disorders and cross-disorder and disorder-specific neural connectivity in two independent populations

  • Tianqi Wang
  • Xiaolong Zhang
  • Ang Li
  • Meifang Zhu
  • Shu Liu
  • Wen Qin
  • Jin Li
  • Chunshui Yu

Major psychiatric disorders, including attention deficit hyperactivity disorder (ADHD), autism (AUT), bipolar disorder (BD), major depressive disorder (MDD), and schizophrenia (SZ), are highly heritable and polygenic. Evidence suggests that these five disorders have both shared and distinct genetic risks and neural connectivity abnormalities. To measure aggregate genetic risks, the polygenic risk score (PGRS) was computed. Two independent general populations (N = 360 and N = 323) were separately examined to investigate whether the cross-disorder PGRS and PGRS for a specific disorder were associated with individual variability in functional connectivity. Consistent altered functional connectivity was found with the bilateral insula: for the left supplementary motor area and the left superior temporal gyrus with the cross-disorder PGRS, for the left insula and right middle and superior temporal lobe associated with the PGRS for autism, for the bilateral midbrain, posterior cingulate, cuneus, and precuneus associated with the PGRS for BD, and for the left angular gyrus and the left dorsolateral prefrontal cortex associated with the PGRS for schizophrenia. No significant functional connectivity was found associated with the PGRS for ADHD and MDD. Our findings indicated that genetic effects on the cross-disorder and disorder-specific neural connectivity of common genetic risk loci are detectable in the general population. Our findings also indicated that polygenic risk contributes to the main neurobiological phenotypes of psychiatric disorders and that identifying cross-disorder and specific functional connectivity related to polygenic risks may elucidate the neural pathways for these disorders.

Details DOI

YNICL Journal 2015 Journal Article

Evoked itch perception is associated with changes in functional brain connectivity

  • Gaëlle Desbordes
  • Ang Li
  • Marco L. Loggia
  • Jieun Kim
  • Peter C. Schalock
  • Ethan Lerner
  • Thanh N. Tran
  • Johannes Ring

Chronic itch, a highly debilitating condition, has received relatively little attention in the neuroimaging literature. Recent studies suggest that brain regions supporting itch in chronic itch patients encompass sensorimotor and salience networks, and corticostriatal circuits involved in motor preparation for scratching. However, how these different brain areas interact with one another in the context of itch is still unknown. We acquired BOLD fMRI scans in 14 atopic dermatitis patients to investigate resting-state functional connectivity before and after allergen-induced itch exacerbated the clinical itch perception in these patients. A seed-based analysis revealed decreased functional connectivity from baseline resting state to the evoked-itch state between several itch-related brain regions, particularly the insular and cingulate cortices and basal ganglia, where decreased connectivity was significantly correlated with increased levels of perceived itch. In contrast, evoked itch increased connectivity between key nodes of the frontoparietal control network (superior parietal lobule and dorsolateral prefrontal cortex), where higher increase in connectivity was correlated with a lesser increase in perceived itch, suggesting that greater interaction between nodes of this executive attention network serves to limit itch sensation via enhanced top-down regulation. Overall, our results provide the first evidence of itch-dependent changes in functional connectivity across multiple brain regions.

Details DOI

IJCAI Conference 2013 Conference Paper

Efficient Learning in Linearly Solvable MDP Models

  • Ang Li
  • Paul R. Schrater

Linearly solvable Markov Decision Process (MDP) models are a powerful subclass of problems with a simple structure that allow the policy to be written directly in terms of the uncontrolled (passive) dynamics of the environment and the goals of the agent. However, there have been no learning algorithms for this class of models. In this research, we develop a robust learning approach to linearly solvable MDPs. To exploit the simple solution for general problems, we show how to construct passive dynamics from any transition matrix, use Bayesian updating to estimate the model parameters and apply approximate and efficient Bayesian exploration to speed learning. In addition, we reduce the computational cost of learning using intermittent Bayesian updating and policy solving. We also gave a polynomial theoretical time complexity bound for the convergence of our learning algorithm, and demonstrate a linear bound for the subclass of the reinforcement learning problems with the property that the transition error depends only on the agent itself. Test results for our algorithm in a grid world are presented, comparing our algorithm with the BEB algorithm. The results showed that our algorithm learned more than the BEB algorithm without losing convergence speed, so that the advantage of our algorithm increased as the environment got more complex. We also showed that our algorithm’s performance is more stable after convergence. Finally, we show how to apply our approach to the Cellular Telephones problem by defining the passive dynamics.

PDF Details DOI
v2026.09.13