Xi Zhang

YNICL Journal 2026 Journal Article

Brain topology alteration in Alzheimer’s disease brain networks: A multi-center study

• Longhao Ma
• Pan Wang
• Dawei Wang
• Hongxiang Yao
• Bo Zhou
• Yonghua Zhao
• Zhengluan Liao
• Yan Chen

Alterations in brain network centrality are key features of Alzheimer's disease (AD) and may offer insights into the disruption of network organization underlying cognitive decline. We introduce a novel centrality metric, DomiRank, to characterize dominance-driven connectivity patterns in the human brain network, using a multi-center MRI dataset comprising 809 participants. Compared with conventional metrics, DomiRank centrality showed greater sensitivity in detecting AD-related network disruptions, particularly within the cingulate gyrus, precuneus, and subcortical hubs such as the basal ganglia-regions critical for cognition. Regional DomiRank alterations were significantly correlated with clinical cognitive scores, indicating their potential relevance to disease severity. Gene enrichment analysis revealed that areas with reduced DomiRank centrality were enriched for genes involved in synaptic signaling and neuronal communication, suggesting molecular mechanisms underlying network vulnerability. These findings highlight DomiRank centrality as a promising biomarker for characterizing network disorganization in AD, linking changes in brain connectivity with underlying molecular processes.

AAAI Conference 2026 Conference Paper

Bridging Vision and Language for Robust Context-Aware Surgical Point Tracking: The VL-SurgPT Dataset and Benchmark

• Rulin Zhou
• Wenlong He
• An Wang
• Jianhang Zhang
• Xuanhui Zeng
• Xi Zhang
• Chaowei Zhu
• Haijun Hu

Accurate point tracking in surgical environments remains challenging due to complex visual conditions, including smoke occlusion, specular reflections, and tissue deformation. While existing surgical tracking datasets provide coordinate information, they lack the semantic context necessary to understand tracking failure mechanisms. We introduce VL-SurgPT, the first large-scale multimodal dataset that bridges visual tracking with textual descriptions of point status in surgical scenes. The dataset comprises 908 in vivo video clips, including 754 for tissue tracking (17,171 annotated points across five challenging scenarios) and 154 for instrument tracking (covering seven instrument types with detailed keypoint annotations). We establish comprehensive benchmarks using eight state-of-the-art tracking methods and propose TG-SurgPT, a text-guided tracking approach that leverages semantic descriptions to improve robustness in visually challenging conditions. Experimental results demonstrate that incorporating point status information significantly improves tracking accuracy and reliability, particularly in adverse visual scenarios where conventional vision-only methods struggle. By bridging visual and linguistic modalities, VL-SurgPT enables the development of context-aware tracking systems crucial for advancing computer-assisted surgery applications that can maintain performance even under challenging intraoperative conditions.

AAAI Conference 2026 Conference Paper

METP: Multi-Granularity Integration of External Covariates for Temporal Point Processes

• Boyang Li
• Lingzheng Zhang
• Fugee Tsung
• Xi Zhang

Accurate modeling of temporal point processes is critical for reliable event forecasting and informed decision-making. While historical event sequences provide a foundation for intensity estimation, existing approaches often neglect external covariates whose lagged effects impact future intensities across multiple temporal granularities. To address this gap, we propose Multi-Granularity Integration of External Covariates for Temporal Point Processes (METP), a framework for incorporating lagged external influences into intensity modeling. METP extracts periodic structures and decomposes external covariate series into multiple temporal granularities. At each granularity, a lag-aware calibration module is introduced to align covariates with event dynamics. Finally, a hierarchical mixture-of-experts strategy is employed to integrate the multi-granular external covariates with historical event embeddings, enabling a representation of the conditional intensity function with enhanced information. Extensive experiments on public and proprietary datasets demonstrate that METP consistently outperforms existing methods in predictive accuracy.

AAAI Conference 2026 Conference Paper

MirrorShield: Towards Dynamic Adaptive Defense Against Jailbreaks via Entropy-Guided Mirror Crafting

• Rui Pu
• Chaozhuo Li
• Rui Ha
• Litian Zhang
• Lirong Qiu
• Xi Zhang

Defending large language models (LLMs) against jailbreak attacks is crucial for ensuring their safe deployment. Existing defense strategies typically rely on predefined static criteria to differentiate between harmful and benign prompts. However, such rigid rules fail to accommodate the inherent complexity and dynamic nature of real-world jailbreak attacks. In this paper, we focus on the novel challenge of adaptive defense against diverse jailbreaks. We propose a new concept "mirror'', which is a dynamically generated prompt that reflects the syntactic structure of the input while ensuring semantic safety. The discrepancies between input prompts and their corresponding mirrors serve as guiding principles for defense. A novel defense model, MirrorShield, is further proposed to detect and calibrate risky inputs based on the crafted mirrors. Evaluated on multiple benchmark datasets and compared against ten state-of-the-art attack methods, MirrorShield demonstrates superior defense performance and promising generalization capabilities.

AAAI Conference 2026 Conference Paper

SecMoE: Communication-Efficient Secure MoE Inference via Select-Then-Compute

• Bowen Shen
• Yuyue Chen
• Peng Yang
• Bin Zhang
• Xi Zhang
• Zoe L. Jiang

Privacy-preserving Transformer inference has gained attention due to the potential leakage of private information. Despite recent progress, existing frameworks still fall short of practical model scales, with gaps up to a hundredfold. A possible way to close this gap is the Mixture of Experts (MoE) architecture, which has emerged as a promising technique to scale up model capacity with minimal overhead. However, given that the current secure two-party (2-PC) protocols allow the server to homomorphically compute the FFN layer with its plaintext model weight, under the MoE setting, this could reveal which expert is activated to the server, exposing token-level privacy about the client's input. While naively evaluating all the experts before selection could protect privacy, it nullifies MoE sparsity and incurs the heavy computational overhead that sparse MoE seeks to avoid. To address the privacy and efficiency limitations above, we propose a 2-PC privacy-preserving inference framework, SecMoE. Unifying per-entry circuits in both the MoE layer and piecewise polynomial functions, SecMoE obliviously selects the extracted parameters from circuits and only computes one encrypted entry, which we refer to as Select-Then-Compute. This makes the model for private inference scale to 63× larger while only having a 15.2× increase in end-to-end runtime. Extensive experiments show that, under 5 expert settings, SecMoE lowers the end-to-end private inference communication by 1.8~7.1× and achieves 1.3~3.8× speedup compared to the state-of-the-art (SOTA) protocols.

AAAI Conference 2026 Conference Paper

TruthfulRAG: Resolving Factual-level Conflicts in Retrieval-Augmented Generation with Knowledge Graphs

• Shuyi Liu
• Yu-Ming Shang
• Xi Zhang

Retrieval-Augmented Generation (RAG) has emerged as a powerful framework for enhancing the capabilities of Large Language Models (LLMs) by integrating retrieval-based methods with generative models. As external knowledge repositories continue to expand and the parametric knowledge within models becomes outdated, a critical challenge for RAG systems is resolving conflicts between retrieved external information and LLMs' internal knowledge, which can significantly compromise the accuracy and reliability of generated content. However, existing approaches to conflict resolution typically operate at the token or semantic level, often leading to fragmented and partial understanding of factual discrepancies between LLMs' knowledge and context, particularly in knowledge-intensive tasks. To address this limitation, we propose TruthfulRAG, the first framework that leverages Knowledge Graphs (KGs) to resolve factual-level knowledge conflicts in RAG systems. Specifically, TruthfulRAG constructs KGs by systematically extracting triples from retrieved content, utilizes query-based graph retrieval to identify relevant knowledge, and employs entropy-based filtering mechanisms to precisely locate conflicting elements and mitigate factual inconsistencies, thereby enabling LLMs to generate faithful and accurate responses. Extensive experiments reveal that TruthfulRAG outperforms existing methods, effectively alleviating knowledge conflicts and improving the robustness and trustworthiness of RAG systems.

YNIMG Journal 2026 Journal Article

Uniformity in happiness and uniqueness in sadness: Naturalistic emotional representation in major depression

• Qingjin Liu
• Xi Zhang
• Jinpeng Niu
• Kangjia Chen
• Jie Xia
• Yaohui He
• Shuo Xu
• Wei Li

Humans develop shared concepts of others' emotions to support adaptive social functioning, yet how these concepts are dynamically represented in major depressive disorder (MDD) during naturalistic movie viewing is not yet fully established. Using functional MRI, we examined patients with MDD (n = 55) and healthy controls (HCs; n = 62) as they freely viewed movie clips depicting happy and sad emotions. Neural similarity was quantified with inter-subject correlation at whole-brain, network, and regional levels, and its association with emotional traits was assessed using inter-subject representational similarity analysis. Compared with HCs, patients with MDD showed significantly reduced whole-brain similarity, particularly during sad contexts. Network analyses revealed that HCs exhibited increased similarity in the limbic network during sadness, reflecting a shared "sadness resonance," whereas patients with higher depressive severity showed widespread disruptions across visual, limbic, dorsal attention, and default mode networks. At the regional level, similarity in the inferior temporal gyrus and lateral occipital cortex was closely linked to individual differences in emotional awareness, with pronounced context- and region-specificity. These findings highlight neural decoupling and heterogeneity as core features of MDD and provide new evidence for potential biomarkers to inform risk assessment and personalized interventions.

TIST Journal 2025 Journal Article

Advancing Session-Based Recommendations with Atten-Mixer+: Dynamic and Adaptive Multi-Level Intent Mining

• Peiyan Zhang
• Jiayan Guo
• Chaozhuo Li
• Liying Kang
• Jaeboum Kim
• Jie Xu
• Xi Zhang
• Yan Zhang

Session-Based Recommendation (SBR) systems, traditionally reliant on complex Graph Neural Networks (GNNs), often face challenges with marginal performance improvements despite increased model complexity. In this article, we dissect the classical GNN-based SBR models and empirically find that the sophisticated GNN propagations might be redundant, given the readout module plays a significant role in GNN-based models. Based on this observation, we introduce Atten-Mixer+, an advanced iteration of our previously developed Multi-Level Attention Mixture Network (Atten-Mixer). Atten-Mixer+ forgoes GNN propagation in favor of a dynamic and adaptive readout process, tailored to the unique characteristics of each session. Different from the vanilla version, Atten-Mixer+ features the Adaptive Intent Scaler (AIS) layer, which dynamically determines the depth of multi-level user intent analysis and a soft allocation approach for generating user intent queries across entire user interaction sequences. This innovative design allows Atten-Mixer+ to capture a nuanced and comprehensive understanding of user behaviors, overcoming the limitations of fixed-length analysis. Empirical evaluations on benchmark datasets highlight Atten-Mixer+’s superior efficiency and effectiveness, marking a significant step forward in the predictive accuracy of SBR systems.

EAAI Journal 2025 Journal Article

Advancing the safety of intelligent rail transit systems: A segmentation network for efficient end-of-track degradation feature extraction

• Tao Ye
• Haoran Chen
• Guopeng Liu
• Liu Liu
• Hongbin Ren
• Xiaosong Li
• Xi Zhang

Accurate segmentation of rail track is crucial for the safe autonomous driving of intelligent trains. Current train operations struggle with insufficient precision in rail track segmentation, primarily due to poor end-of-track segmentation performance caused by degradation of track-end features. To address these challenges, we propose Rail Track End Wise Network (RTEW-Net), an effective rail track end wise segmentation method. This network utilizes Full-Transformer Module (FTM) for effective track feature extraction and integrates the Global Response Normalization (GRN) module to handle drastic lighting changes. Additionally, we designed the Wise Weigh Maintain (WWM) method to enhance feature learning and retain track features. To validate its effectiveness, we constructed the RailMixed2024 (RM2024) dataset. Our model achieves high-precision global rail track segmentation and optimizes end detection. Experimental results demonstrate that RTEW-Net exhibits outstanding performance on the RM2024 and RailSem19 datasets, establishing it as the state-of-the-art (SOTA) in this field.

IJCAI Conference 2025 Conference Paper

Automated Detection of Pre-training Text in Black-box LLMs

• Ruihan Hu
• Yu-Ming Shang
• Jiankun Peng
• Wei Luo
• Yazhe Wang
• Xi Zhang

Detecting whether a given text is a member in the pre-training data of Large Language Models (LLMs) is crucial for ensuring data privacy and copyright protection. Most existing methods rely on the LLM's hidden information (e. g. , model parameters or token probabilities), making them ineffective in the black-box setting, where only input and output texts are accessible. Although some methods have been proposed for the black-box setting, they rely on massive manual efforts such as designing complicated questions or instructions. To address these issues, we propose VeilProbe, the first framework for automatically detecting LLMs' pre-training texts in a black-box setting without human intervention. VeilProbe utilizes a sequence-to-sequence mapping model to infer the latent mapping feature between the input text and the corresponding output suffix generated by the LLM. Then it performs the key token perturbations to obtain more distinguishable membership features. Additionally, considering real-world scenarios where the ground-truth training text samples are limited, a prototype-based membership classifier is introduced to alleviate the overfitting issue. Extensive evaluations on three widely used datasets demonstrate that our framework is effective and superior in the black-box setting.

NeurIPS Conference 2025 Conference Paper

BADiff: Bandwidth Adaptive Diffusion Model

• Xi Zhang
• Hanwei Zhu
• Yan Zhong
• Jiamang Wang
• Weisi Lin

In this work, we propose a novel framework to enable diffusion models to adapt their generation quality based on real-time network bandwidth constraints. Traditional diffusion models produce high-fidelity images by performing a fixed number of denoising steps, regardless of downstream transmission limitations. However, in practical cloud-to-device scenarios, limited bandwidth often necessitates heavy compression, leading to loss of fine textures and wasted computation. To address this, we introduce a joint end-to-end training strategy where the diffusion model is conditioned on a target quality level derived from the available bandwidth. During training, the model learns to adaptively modulate the denoising process, enabling early-stop sampling that maintains perceptual quality appropriate to the target transmission condition. Our method requires minimal architectural changes and leverages a lightweight quality embedding to guide the denoising trajectory. Experimental results demonstrate that our approach significantly improves the visual fidelity of bandwidth-adapted generations compared to naive early-stopping, offering a promising solution for efficient image delivery in bandwidth-constrained environments. Code is available at: https: //github. com/xzhang9308/BADiff.

AAAI Conference 2025 Conference Paper

Beyond Text: Fine-Grained Multi-Modal Fact Verification with Hypergraph Transformers

• Hui Pang
• Chaozhuo Li
• Litian Zhang
• Senzhang Wang
• Xi Zhang

Fact verification has become increasingly vital in the internet age, driven by the proliferation of false claims and political misinformation. While traditional methods rely predominantly on text-based evidence, multi-modal evidence introduces richer sources of information, offering valuable insights for claim verification. Existing multi-modal verification models often focus on superficial correlations between claims and evidence, neglecting the complex semantic interactions present in fine-grained multi-modal signals. In this paper, we propose a novel framework for multi-modal fact-checking, named Hypergraph Transformer-based Multi-modal Fact-Checking (HGTMFC). Our approach captures high-order relationships between different modalities of evidence and claims by leveraging hypergraphs. HGTMFC models the intricate relationships among evidence across various modalities and enhances information propagation through a transformer-based mechanism embedded within the hypergraph. Moreover, we utilize linegraphs to refine this propagation process, further strengthening the model's reasoning capabilities. Experiments on benchmark datasets demonstrate that our model significantly outperforms existing approaches in multi-modal fact verification.

IJCAI Conference 2025 Conference Paper

Curriculum Hierarchical Knowledge Distillation for Bias-Free Survival Prediction

• Chaozhuo Li
• Zhihao Tang
• Mingji Zhang
• Zhiquan Liu
• Litian Zhang
• Xi Zhang

Survival prediction is a pivotal task for estimating mortality risk within a given timeframe based on whole slide images (WSIs). Conventional models typically assume that WSIs across patients are independent and identically distributed, an assumption that may not hold due to inherent variability in WSI preparation and the uncertain condition of infected tissues. These uncontrollable external factors introduce significant variability in the numbers and resolutions of WSIs across patients, leading to bias and compromised performance, particularly for tail patients with limited data. In this paper, we propose a novel approach, PathoKD, based on knowledge distillation. Recognizing the hierarchical nature of disease progression and the data scarcity issues associated with vanilla knowledge distillation methods, PathoKD integrates a novel curriculum learning framework with hierarchical knowledge distillation. This integration effectively mitigates the performance gap between head and tail patients, thereby enhancing prediction accuracy across patient groups. Our proposal is extensively evaluated over popular datasets and experimental results demonstrate its superiority.

AAAI Conference 2025 Conference Paper

Deep Rank-One Tensor Functional Factorization for Multi-Dimensional Data Recovery

• Yanyi Li
• Xi Zhang
• Yisi Luo
• Deyu Meng

Many real-world data are inherently multi-dimensional, e.g., color images, videos, and hyperspectral images. How to effectively and compactly represent these multi-dimensional data within a unified framework is an important pursuit. Previous methods focus on tensor factorizations, convolutional networks, or diffusion models for multi-dimensional data representation, which may not fully utilize inherent data structures and may lead to redundant parameters. In this work, we propose a Deep Rank-One Tensor Functional Factorization (DRO-TFF), which internally utilizes more comprehensive data priors facilitated by much fewer parameters. Concretely, our DRO-TFF consists of three organically integrated blocks: compact rank-one factorizations in the spatial domain, a deep transform to capture underlying low-dimensional structures, and smooth factors parameterized by implicit neural representations. Through a series of theoretical analysis, we show the rich data priors encoded in the DRO-TFF structure, e.g., Lipschitz smoothness and low-rankness. Extensive experiments on multi-dimensional data recovery problems, such as image and video inpainting, image denoising, and hyperspectral mixed noise removal, showcase the effectiveness of the proposed method.

IJCAI Conference 2025 Conference Paper

Diffusion Guided Propagation Augmentation for Popularity Prediction

• Chaozhuo Li
• Tianqi Yang
• Litian Zhang
• Xi Zhang

The prediction of information popularity propagation is critical for applications such as recommendation systems, targeted advertising, and social media trend analysis. Traditional approaches primarily rely on historical cascade data, often sacrificing timeliness for prediction accuracy. These methods capture aggregate diffusion patterns but fail to account for the complex temporal dynamics of early-stage propagation. In this paper, we introduce Diffusion Guided Propagation Augmentation(DGPA), a novel framework designed to improve early-stage popularity prediction. DGPA models cascade dynamics by leveraging a generative approach, where a temporal conditional interpolator serves as a noising process and forecasting as a denoising process. By iteratively generating cascade representations through a sampling procedure, DGPA effectively incorporates the evolving time steps of diffusion, significantly enhancing prediction timeliness and accuracy. Extensive experiments on benchmark datasets from Twitter, Weibo, and APS demonstrate that DGPA outperforms state-of-the-art methods in early-stage popularity prediction.

JBHI Journal 2025 Journal Article

Disentangled Representation Learning for Capturing Individualized Brain Atrophy via Pseudo-Healthy Synthesis

• Zhuangzhuang Li
• Kun Zhao
• Pindong Chen
• Dawei Wang
• Hongxiang Yao
• Bo Zhou
• Jie Lu
• Pan Wang

Brain atrophy emerges as a distinctive hallmark in various neurodegenerative diseases, demonstrating a progressive trajectory across diverse disease stages and concurrently manifesting in tandem with a discernible decline in cognitive abilities. Understanding the individualized patterns of brain atrophy is critical for precision medicine and the prognosis of neurodegenerative diseases. However, it is difficult to obtain longitudinal data to compare changes before and after the onset of diseases. In this study, we present a deep disentangled generative model (DDGM) for capturing individualized atrophy patterns via disentangling patient images into “realistic” healthy counterfactual images and abnormal residual maps. The proposed DDGM consists of four modules: normal MRI synthesis, residual map synthesis, input reconstruction module, and mutual information neural estimator (MINE). The MINE and adversarial learning strategy together ensure independence between disease-related features and features shared by both disease and healthy controls. In addition, we proposed a comprehensive evaluation of the effectiveness of synthetic pseudo-healthy images, focusing on both their healthiness and subject identity. The results indicated that the proposed DDGM effectively preserves these characteristics in the synthesized pseudo-healthy images, outperforming existing methods. The proposed method demonstrates robust generalization capabilities across two independent datasets from different races and sites. Analysis of the disease residual/saliency maps revealed specific atrophy patterns associated with Alzheimer's disease (AD), particularly in the hippocampus and amygdala regions. These accurate individualized atrophy patterns enhance the performance of AD classification tasks, resulting in an improvement in classification accuracy to 92. 50 $\pm$ 2. 70%.

YNIMG Journal 2025 Journal Article

Early wounds, delayed consequences: Brain-behavior modeling reveals neural pathways linking childhood trauma to procrastination

• Luo Xu
• Yao Yin
• Xueke Wang
• Ting Xu
• Xi Zhang
• Tingyong Feng

Childhood trauma has enduring effects on emotional and cognitive functioning, yet its impact on procrastination, particularly from a neurodevelopmental perspective, remains poorly understood. To achieve this, we employed resting-state functional MRI in conjunction with standardized behavioral assessments of childhood trauma, trait anxiety, self-control, and procrastination across two datasets (discovery dataset: n = 760; validation dataset: n = 429). By leveraging the advanced predictive analytics-including connectome-based predictive modeling (CPM) and least absolute shrinkage and selection operator (LASSO) regression-we aimed to elucidate the neural basis linking childhood trauma to procrastination. Our behavioral results revealed that childhood trauma was a significant predictor of elevated procrastination tendencies, with this association mediated by increased trait anxiety and reduced self-control. At the neural level, the predictive modeling using CPM and LASSO regression demonstrated that functional connectivity within and between the frontoparietal network (FPN), salience network (SAN), visual network (VN), and cerebellum significantly predicted childhood trauma. These patterns likely reflect trauma-related disruptions in higher-order cognitive control (e.g., self-control) and increased affective reactivity (e.g., trait anxiety). More importantly, the mediation analyses further confirmed that trait anxiety and self-control jointly mediate the relationship between trauma-related neural network connectivity and procrastination. These findings presented novel evidence that childhood trauma is associated with procrastination via functional alterations in large-scale neural networks implicated in self-control and emotion regulation, providing critical insights into the long-term behavioral consequences of early-life adversity, and informing the development of targeted interventions to reduce procrastination in trauma-exposed individuals.

EAAI Journal 2025 Journal Article

Erratum to “Advancing the safety of intelligent rail transit systems: A segmentation network for efficient end-of-track degradation feature extraction” [Eng. Appl. Artif. Intellig. 159 (2025) 111527]

• Tao Ye
• Haoran Chen
• Guopeng Liu
• Liu Liu
• Hongbin Ren
• Xiaosong Li
• Xi Zhang

AAAI Conference 2025 Conference Paper

From Representation Space to Prognostic Insights: Whole Slide Image Generation with Hierarchical Diffusion Model for Survival Prediction

• Zhihao Tang
• Xi Zhang
• Chaozhuo Li

Deep learning has significantly enhanced survival prediction using whole slide images (WSIs) by adopting a two-stage learning paradigm: WSI preparation and patient-level prediction. While existing research generally concentrates on developing advanced patient-level prediction modules, the critical importance of WSI preparation has been largely overlooked. In practice, WSI preparation is influenced by numerous factors, including tissue heterogeneity, sampling strategies, and technical considerations. These uncontrollable external factors incur variability in the number of WSIs among patients, introducing significant bias and resulting in inferior performance for patients with few WSIs. To address this challenge, we propose a novel approach named WSI-Diffusion. Unlike existing WSI generation models that produce augmented versions of input WSIs, our method generates entirely new WSIs in representation space to serve as complementary data. WSIDiffusion employs a two-stage hierarchical diffusion process. Two novel modules, WSI-level and patch-level Diffusers are designed to capture complex correlations between WSIs and patches. The generated WSIs are integrated as supplementary data, and a light patient-level prediction module is then trained for survival prediction. Experimental results across five datasets demonstrate the superiority of our proposal.

EAAI Journal 2025 Journal Article

Generalized zero-shot fault diagnosis method for chillers based on cross-modal information compression

• Kexin Jiang
• Xuejin Gao
• Huayun Han
• Huihui Gao
• Yongsheng Qi
• Xi Zhang
• Liang Zhao

In practical chiller fault diagnosis applications, target fault training data is often inaccessible, which poses significant challenges to the effectiveness of data-driven diagnostic approaches. Generalized zero-shot fault diagnosis (GZSFD) aims to detect and classify all fault types without relying on fault samples from all categories during training. GZSFD models are trained exclusively on data from seen faults—those with available historical data—making them prone to bias toward these seen classes during inference. However, existing methods fail to exploit the fine-grained semantics of fault attributes, resulting in suboptimal performance when addressing feature bias toward seen classes. In this article, an innovative GZSFD framework for chillers based on cross-modal information compression is proposed to overcome this difficulty. This method constructs a cross-modal feature fusion attention (CmFFA) and information compression module, and integrates them into a variational autoencoder with generative adversarial network (VAEGAN), and establishes a CmFFA-VAEGAN network to synthesize high-quality unseen fault samples. Specifically, the CmFFA module effectively aligns local regions of the virtual samples with key words in the textual attributes, enabling the model to attend to fine-grained semantic information and robustly fuse features from different modalities. This facilitates the generation of virtual samples that more accurately reflect the characteristics of the target modality. In addition, an information bottleneck (IB) layer is introduced at the output of the generator to compress redundant information within the fused features, retaining only the key information most relevant to the data augmentation task. This design enhances cross-modal consistency between the synthesized samples and their corresponding attributes, while alleviating the feature shift of unseen class samples toward seen categories. Extensive experiments are designed and executed on the chiller dataset. Experimental results demonstrate that the proposed framework effectively mitigates the bias of synthesized unseen fault samples toward seen categories, leading to a significant improvement in diagnostic accuracy.

AAAI Conference 2025 Conference Paper

Ghidorah: Towards Robust Multi-Scale Information Diffusion Prediction via Test-Time Training

• Wenting Zhu
• Chaozhuo Li
• Litian Zhang
• Senzhang Wang
• Xi Zhang

Information diffusion prediction (IDP) is a pivotal task for understanding the dynamics of information propagation within social networks. Conventional models typically adhere to a fixed learning-based paradigm, where the trained prediction model remains static during the inference phase. This paradigm presupposes that the data is independent and identically distributed, an assumption that may not hold true due to the inherently open nature of social media and the uncertainty and variability in user behavior. In this paper, we address the novel problem of out-of-distribution (OOD) shifts within IDP tasks and propose a new test-time training-based model for multi-scale IDP tasks, named Ghidorah. Our approach focuses on adapting a subset of model parameters to accommodate the unique characteristics of test samples through self-supervised learning (SSL) tasks. Ghidorah comprises three components: the macroscopic prediction branch, the microscopic prediction branch, and the auxiliary SSL branch. The auxiliary SSL task employs a masked autoencoder-based loss to fine-tune the model for specific test samples prior to prediction. Furthermore, Ghidorah integrates invariant learning to capture robust representations while mitigating spurious correlations. To our knowledge, Ghidorah is the first work to introduce a test-time training framework specifically designed to address the critical yet often overlooked OOD challenges in IDP. Experimental results across several benchmark datasets validate the superiority of our approach.

NeurIPS Conference 2025 Conference Paper

Learning Grouped Lattice Vector Quantizers for Low-Bit LLM Compression

• Xi Zhang
• Xiaolin Wu
• Jiamang Wang
• Weisi Lin

Large Language Models (LLMs) have demonstrated remarkable capabilities but typically require extensive computational resources and memory for inference. Post-training quantization (PTQ) can effectively reduce these demands by storing weights in lower bit-width formats. However, standard uniform quantization often leads to notable performance degradation, particularly in low-bit scenarios. In this work, we introduce a Grouped Lattice Vector Quantization (GLVQ) framework that assigns each group of weights a customized lattice codebook, defined by a learnable generation matrix. To address the non-differentiability of the quantization process, we adopt Babai rounding to approximate nearest-lattice-point search during training, which enables stable optimization of the generation matrices. Once trained, decoding reduces to a simple matrix-vector multiplication, yielding an efficient and practical quantization pipeline. Experiments on multiple benchmarks show that our approach achieves a better trade-off between model size and accuracy compared to existing post-training quantization baselines, highlighting its effectiveness in deploying large models under stringent resource constraints. Our source code is available on GitHub repository: https: //github. com/xzhang9308/GLVQ.

NeurIPS Conference 2025 Conference Paper

Look Before You Leap: A GUI-Critic-R1 Model for Pre-Operative Error Diagnosis in GUI Automation

• Yuyang Wanyan
• Xi Zhang
• Haiyang Xu
• Haowei Liu
• Junyang Wang
• Jiabo Ye
• Yutong Kou
• Ming Yan

In recent years, Multimodal Large Language Models (MLLMs) have been extensively utilized for multimodal reasoning tasks, including Graphical User Interface (GUI) automation. Unlike general offline multimodal tasks, GUI automation is executed in online interactive environments, necessitating step-by-step decision-making based on the real-time status of the environment. This task has a lower tolerance for decision-making errors at each step, as any mistakes may cumulatively disrupt the process and potentially lead to irreversible outcomes like deletions or payments. To address these issues, we introduce a pre-operative critic mechanism that provides effective feedback prior to the actual execution, by reasoning about the potential outcome and correctness of actions. Specifically, we propose a Suggestion-aware Group Relative Policy Optimization (S-GRPO) strategy to construct our pre-operative critic model GUI-Critic-R1, incorporating a novel suggestion reward to enhance the reliability of the model's feedback. Furthermore, we develop a reasoning-bootstrapping based data collection pipeline to create a GUI-Critic-Train and a GUI-Critic-Test, filling existing gaps in GUI critic data. Static experiments on the GUI-Critic-Test across both mobile and web domains reveal that our GUI-Critic-R1 offers significant advantages in critic accuracy compared to current MLLMs. Dynamic evaluation on GUI automation benchmark further highlights the effectiveness and superiority of our model, as evidenced by improved success rates and operational efficiency. The code is available at https: //github. com/X-PLUG/MobileAgent/tree/main/GUI-Critic-R1.

ICLR Conference 2025 Conference Paper

Meta Flow Matching: Integrating Vector Fields on the Wasserstein Manifold

• Lazar Atanackovic
• Xi Zhang
• Brandon Amos
• Mathieu Blanchette
• Leo J. Lee
• Yoshua Bengio
• Alexander Tong 0001
• Kirill Neklyudov

Numerous biological and physical processes can be modeled as systems of interacting entities evolving continuously over time, e.g. the dynamics of communicating cells or physical particles. Learning the dynamics of such systems is essential for predicting the temporal evolution of populations across novel samples and unseen environments. Flow-based models allow for learning these dynamics at the population level - they model the evolution of the entire distribution of samples. However, current flow-based models are limited to a single initial population and a set of predefined conditions which describe different dynamics. We argue that multiple processes in natural sciences have to be represented as vector fields on the Wasserstein manifold of probability densities. That is, the change of the population at any moment in time depends on the population itself due to the interactions between samples. In particular, this is crucial for personalized medicine where the development of diseases and their respective treatment response depend on the microenvironment of cells specific to each patient. We propose *Meta Flow Matching* (MFM), a practical approach to integrate along these vector fields on the Wasserstein manifold by amortizing the flow model over the initial populations. Namely, we embed the population of samples using a Graph Neural Network (GNN) and use these embeddings to train a Flow Matching model. This gives MFM the ability to generalize over the initial distributions, unlike previously proposed methods. We demonstrate the ability of MFM to improve the prediction of individual treatment responses on a large-scale multi-patient single-cell drug screen dataset.

AAAI Conference 2025 Conference Paper

MRR-FV: Unlocking Complex Fact Verification with Multi-Hop Retrieval and Reasoning

• Liwen Zheng
• Chaozhuo Li
• Litian Zhang
• Haoran Jia
• Senzhang Wang
• Zheng Liu
• Xi Zhang

The pervasive spread of misinformation on social networks highlights the critical necessity for effective fact verification systems. Traditional approaches primarily focus on pairwise correlations between claims and evidence, often neglecting comprehensive multi-hop retrieval and reasoning, which results in suboptimal performance when dealing with complex claims. In this paper, we propose MRR-FV, a generative retrieval-enhanced model designed to address the novel challenge of Multi-hop Retrieval and Reasoning for Fact Verification, which integrates two core modules: Generative Multi-hop Retriever and the Hierarchical Interaction Reasoner. MRR-FV utilizes an autoregressive model for iterative multi-hop evidence retrieval, complemented by a pre-trained compressor to address the challenge of intention shift across retrieval hops. For claim verification, we propose a hierarchical interaction reasoner that conducts intra-sentence reasoning to capture long-term semantic dependencies and inter-sentence reasoning across multi-hop evidence subgraphs to reveal complex evidence interactions. Experimental evaluations on the FEVER and HOVER datasets demonstrate the superior performance of our model in both claim verification and evidence retrieval tasks.

EAAI Journal 2025 Journal Article

Multi-action deep reinforcement learning-based collaborative production dynamic scheduling approach for networked multi-factory under mass customization

• Mingzhou Liu
• Zhengfu Ding
• Xi Zhang
• Lin Ling
• Maogen Ge

Increasing customization demand is accelerating the decentralization of manufacturing systems, elevating the distributed scheduling problem (DSP) to a prominent research focus. However, the dynamic characteristics of mass customization (MC) and the practically adaptable networked multi-factory systems remain underexplored. To address this gap, this study formulates a networked multi-factory collaborative production scheduling problem in the context of MC (MC-NMF-CPSP), and proposes a solution framework based on multi-action deep reinforcement learning (MA-DRL). The framework formulates a multiple Markov Decision Process (MMDP) incorporating a multi-dimensional state space, a two-stage action space, and a composite reward, providing a unified modeling paradigm for high-dimensional complex scheduling problems. To accurately characterize the dynamic demand, a consistent measurement model is proposed and embedded within the state space for efficient state representation. To enhance learning performance, an improved Proximal Policy Optimization (PPO) is developed, and a Multi-PPO architecture is integrated to construct a networked multi-factory collaborative production scheduling model (MaDRL-NMF-CPSM), which demonstrates exceptional coordination and adaptability under dynamic demand environments. Furthermore, a retraining mechanism is introduced to enable continual model evolution, thereby improving robustness and generalization. Extensive experiments show that the proposed method significantly outperforms several benchmark algorithms and heuristic rules in terms of convergence efficiency, scheduling performance, and dynamic adaptability, offering a novel theoretical foundation and practical solution for collaborative scheduling in networked manufacturing systems under dynamic demand.

EAAI Journal 2025 Journal Article

Multi-regularized tensor-based framework for identifying hard landings

• Chenyang Chang
• Yu An
• Xi Zhang

Hard landings are a significant concern in civil aviation, often resulting in aircraft structural damage, financial losses, and compromised passenger safety. Automating the detection of such incidents faces challenges due to the complexities of Quick Access Recorder (QAR) data, which exhibit multi-channel interdependencies and temporal dynamics. Furthermore, environmental factors tied to flight-variant data, such as the geographic attributes of landing airports, can influence the occurrence of hard landings, yet these factors are often neglected in existing methodologies. This omission limits the practical utility of current approaches for enhancing safety in civil aviation. To address these challenges, we propose a multi-regularized tensor-based framework that models QAR data as a high-order tensor and applies tensor decomposition to extract latent patterns that characterize hard landing scenarios. The model incorporates tailored regularization terms to address both temporal correlations and inter-channel couplings across aircraft systems. To enable efficient computation, we develop a customized Block Coordinate Descent (BCD) algorithm, designed for efficient processing with high-dimensional factor matrices. The effectiveness of the proposed framework is validated using real-world civil aviation data from China, demonstrating superior performance in identifying hard landings.

NeurIPS Conference 2025 Conference Paper

One SPACE to Rule Them All: Jointly Mitigating Factuality and Faithfulness Hallucinations in LLMs

• Pengbo Wang
• Chaozhuo Li
• Chenxu Wang
• Liwen Zheng
• Litian Zhang
• Xi Zhang

LLMs have demonstrated unprecedented capabilities in natural language processing, yet their practical deployment remains hindered by persistent factuality and faithfulness hallucinations. While existing methods address these hallucination types independently, they inadvertently induce performance trade-offs, as interventions targeting one type often exacerbate the other. Through empirical and theoretical analysis of activation space dynamics in LLMs, we reveal that these hallucination categories share overlapping subspaces within neural representations, presenting an opportunity for concurrent mitigation. To harness this insight, we propose SPACE, a unified framework that jointly enhances factuality and faithfulness by editing shared activation subspaces. SPACE establishes a geometric foundation for shared subspace existence through dual-task feature modeling, then identifies and edits these subspaces via a hybrid probe strategy combining spectral clustering and attention head saliency scoring. Experimental results across multiple benchmark datasets demonstrate the superiority of our approach.

NeurIPS Conference 2025 Conference Paper

Online Functional Tensor Decomposition via Continual Learning for Streaming Data Completion

• Xi Zhang
• Yanyi Li
• Yisi Luo
• Qi Xie
• Deyu Meng

Online tensor decompositions are powerful and proven techniques that address the challenges in processing high-velocity streaming tensor data, such as traffic flow and weather system. The main aim of this work is to propose a novel online functional tensor decomposition (OFTD) framework, which represents a spatial-temporal continuous function using the CP tensor decomposition parameterized by coordinate-based implicit neural representations (INRs). The INRs allow for natural characterization of continually expanded streaming data by simply adding new coordinates into the network. Particularly, our method transforms the classical online tensor decomposition algorithm into a more dynamic continual learning paradigm of updating the INR weights to fit the new data without forgetting the previous tensor knowledge. To this end, we introduce a long-tail memory replay method that adapts to the local continuity property of INR. Extensive experiments for streaming tensor completion using traffic, weather, user-item, and video data verify the effectiveness of the OFTD approach for streaming data analysis. This endeavor serves as a pivotal inspiration for future research to connect classical online tensor tools with continual learning paradigms to better explore knowledge underlying streaming tensor data.

NeurIPS Conference 2025 Conference Paper

Robust Explanations of Graph Neural Networks via Graph Curvatures

• Yazheng Liu
• Xi Zhang
• Sihong Xie
• Hui Xiong

Explaining graph neural networks (GNNs) is a key approach to improve the trustworthiness of GNN in high-stakes applications, such as finance and healthcare. However, existing methods are vulnerable to perturbations, raising concerns about explanation reliability. Prior methods enhance explanation robustness using model retraining or explanation ensemble, with certain weaknesses. Retraining leads to models that are different from the original target model and misleading explanations, while ensemble can produce contradictory results due to different inputs or models. To improve explanation robustness without the above weaknesses, we take an unexplored route and exploit the two edge geometry properties curvature and resistance to enhance explanation robustness. We are the first to prove that these geometric notions can be used to bound explanation robustness. We design a general optimization algorithm to incorporate these geometric properties into a wide spectrum of base GNN explanation methods to enhance the robustness of base explanations. We empirically show that our method outperforms six base explanation methods in robustness across nine datasets spanning node classification, link prediction, and graph classification tasks, improving fidelity in 80\% of the cases and achieving up to a 10\% relative improvement in robust performance. The code is available at https: //github. com/yazhengliu/Robust explanation curvature.

YNIMG Journal 2025 Journal Article

The analyses of structural covariance and structural covariance similarity of cortical morphological measures

• Xi Zhang
• Yanmin Peng
• Dongyue Li
• Ailin Hou
• Meng Liang
• Chunshui Yu

Structural covariance refers to the concurrent changes in one morphological measure between two brain regions. Structural covariance of cortical morphological measures such as cortical thickness (CT), surface area (SA), and cortical volume (CV) have been applied to identify brain structural differences between patients with neuropsychiatric disorders and healthy controls. However, the precise relationships between structural covariance patterns of different cortical measures remain largely unknown. Here, we optimized the preprocessing and calculation approaches of structural covariances and investigated both global (whole-brain-level) and regional (brain-region-level) structural covariance similarities between CT, SA, and CV in 35,580 individuals. We found that Pearson correlation outperformed partial correlation due to generating fewer negative correlations of uncertain biological significance and principal component regression outperformed the regressions of total intracranial volume and respective global measures in removing global effects and reducing negative correlations. We observed that both global and regional covariance similarities of SA-CV were much higher than those of CT-CV and CT-SA, although they were influenced by the selection of atlases and covariance values. We also found age and sex effects on structural covariances and age effects on covariance similarities. The higher SA-CV covariance similarities than CT-CV indicates that SA contributes more to CV covariance than CT, although CV is derived from both CT and SA. The lack of CT-SA covariance similarities suggests that CT and SA have different covariance patterns and should be used in combination in structural covariance studies.

AAAI Conference 2025 Conference Paper

When Open-Vocabulary Visual Question Answering Meets Causal Adapter: Benchmark and Approach

• Feifei Zhang
• Zhaoyi Zhang
• Xi Zhang
• Changsheng Xu

Visual Question Answering (VQA) is a multifaceted task that integrates computer vision and natural language processing to produce textual answers from images and questions. Existing VQA benchmarks predominantly adhere to a closed-set paradigm, limiting their ability to address arbitrary, unseen answers, and thus falling short in real-world scenarios. To address this limitation, we introduce the Open-Vocabulary Visual Question Answering (OVVQA) benchmark, specifically designed to evaluate models under open-world conditions by assessing their performance on both base classes (seen, common answers) and novel classes (unseen, rare answers). In conjunction with this benchmark, we propose a model-agnostic Causal Adapter to combat the inherent bias found in current VQA tasks. Our approach leverages front-door adjustment to enhance causal reasoning, significantly improving model performance on novel categories while maintaining accuracy on base classes. Additionally, we introduce an adaptive transfer loss to facilitate the transfer of more knowledge from the pretrained model to our OVVQA task. Extensive experiments across multiple datasets validate the superiority of our method over existing state-of-the-art approaches, demonstrating its robust generalization and adaptability in open-world VQA scenarios.

YNIMG Journal 2024 Journal Article

Brain fingerprinting and cognitive behavior predicting using functional connectome of high inter-subject variability

• Jiayu Lu
• Tianyi Yan
• Lan Yang
• Xi Zhang
• Jiaxin Li
• Dandan Li
• Jie Xiang
• Bin Wang

The functional connectivity (FC) graph of the brain has been widely recognized as a ``fingerprint'' that can be used to identify individuals from a group of subjects. Research has indicated that individual identification accuracy can be improved by eliminating the impact of shared information among individuals. However, current research extracts not only shared information of inter-subject but also individual-specific information from FC graphs, resulting in incomplete separation of shared information and fingerprint information among individuals, leading to lower individual identification accuracy across all functional magnetic resonance imaging (fMRI) states session pairs and poor cognitive behavior prediction performance. In this paper, we propose a method to enhance inter-subject variability combining conditional variational autoencoder (CVAE) network and sparse dictionary learning (SDL) module. By embedding fMRI state information in the encoding and decoding processes, the CVAE network can better capture and represent the common features among individuals and enhance inter-subject variability by residual. Our experimental results on Human Connectome Project (HCP) data show that the refined connectomes obtained by using CVAE with SDL can accurately distinguish an individual from the remaining participants. The success accuracies reached 99.7 % and 99.6 % in the session pair rest1-rest2 and reverse rest2-rest1, respectively. In the identification experiment involving task-task combinations carried out on the same day, the identification accuracies ranged from 94.2 % to 98.8 %. Furthermore, we showed the Frontoparietal and Default networks make the most significant contributions to individual identification and the edges that significantly contribute to individual identification are found within and between the Frontoparietal and Default networks. Additionally, high-level cognitive behaviors can also be better predicted with the obtained refined connectomes, suggesting that higher fingerprinting can be useful for resulting in higher behavioral associations. In summary, our proposed framework provides a promising approach to use functional connectivity networks for studying cognition and behavior, promoting a deeper understanding of brain functions.

NeurIPS Conference 2024 Conference Paper

Enhancing Robustness of Graph Neural Networks on Social Media with Explainable Inverse Reinforcement Learning

• Yuefei Lyu
• Chaozhuo Li
• Sihong Xie
• Xi Zhang

Adversarial attacks against graph neural networks (GNNs) through perturbations of the graph structure are increasingly common in social network tasks like rumor detection. Social media platforms capture diverse attack sequence samples through both machine and manual screening processes. Investigating effective ways to leverage these adversarial samples to enhance robustness is imperative. We improve the maximum entropy inverse reinforcement learning (IRL) method with the mixture-of-experts approach to address multi-source graph adversarial attacks. This method reconstructs the attack policy, integrating various attack models and providing feature-level explanations, subsequently generating additional adversarial samples to fortify the robustness of detection models. We develop precise sample guidance and a bidirectional update mechanism to reduce the deviation caused by imprecise feature representation and negative sampling within the large action space of social graphs, while also accelerating policy learning. We take rumor detector as an example targeted GNN model on real-world rumor datasets. By utilizing a small subset of samples generated by various graph adversarial attack methods, we reconstruct the attack policy, closely approximating the performance of the original attack method. We validate that samples generated by the learned policy enhance model robustness through adversarial training and data augmentation.

NeurIPS Conference 2024 Conference Paper

Learning Optimal Lattice Vector Quantizers for End-to-end Neural Image Compression

• Xi Zhang
• Xiaolin Wu

It is customary to deploy uniform scalar quantization in the end-to-end optimized Neural image compression methods, instead of more powerful vector quantization, due to the high complexity of the latter. Lattice vector quantization (LVQ), on the other hand, presents a compelling alternative, which can exploit inter-feature dependencies more effectively while keeping computational efficiency almost the same as scalar quantization. However, traditional LVQ structures are designed/optimized for uniform source distributions, hence nonadaptive and suboptimal for real source distributions of latent code space for Neural image compression tasks. In this paper, we propose a novel learning method to overcome this weakness by designing the rate-distortion optimal lattice vector quantization (OLVQ) codebooks with respect to the sample statistics of the latent features to be compressed. By being able to better fit the LVQ structures to any given latent sample distribution, the proposed OLVQ method improves the rate-distortion performances of the existing quantization schemes in neural image compression significantly, while retaining the amenability of uniform scalar quantization.

NeurIPS Conference 2024 Conference Paper

Linear Uncertainty Quantification of Graphical Model Inference

• Chenghua Guo
• Han Yu
• Jiaxin Liu
• Chao Chen
• Qi Li
• Sihong Xie
• Xi Zhang

Uncertainty Quantification (UQ) is vital for decision makers as it offers insights into the potential reliability of data and model, enabling more informed and risk-aware decision-making. Graphical models, capable of representing data with complex dependencies, are widely used across domains. Existing sampling-based UQ methods are unbiased but cannot guarantee convergence and are time-consuming on large-scale graphs. There are fast UQ methods for graphical models with closed-form solutions and convergence guarantee but with uncertainty underestimation. We propose LinUProp, a UQ method that utilizes a novel linear propagation of uncertainty to model uncertainty among related nodes additively instead of multiplicatively, to offer linear scalability, guaranteed convergence, and closed-form solutions without underestimating uncertainty. Theoretically, we decompose the expected prediction error of the graphical model and prove that the uncertainty computed by LinUProp is the generalized variance component of the decomposition. Experimentally, we demonstrate that LinUProp is consistent with the sampling-based method but with linear scalability and fast convergence. Moreover, LinUProp outperforms competitors in uncertainty-based active learning on four real-world graph datasets, achieving higher accuracy with a lower labeling budget.

NeurIPS Conference 2024 Conference Paper

Mobile-Agent-v2: Mobile Device Operation Assistant with Effective Navigation via Multi-Agent Collaboration

• Junyang Wang
• Haiyang Xu
• Haitao Jia
• Xi Zhang
• Ming Yan
• Weizhou Shen
• Ji Zhang
• Fei Huang

Mobile device operation tasks are increasingly becoming a popular multi-modal AI application scenario. Current Multi-modal Large Language Models (MLLMs), constrained by their training data, lack the capability to function effectively as operation assistants. Instead, MLLM-based agents, which enhance capabilities through tool invocation, are gradually being applied to this scenario. However, the two major navigation challenges in mobile device operation tasks — task progress navigation and focus content navigation — are difficult to effectively solve under the single-agent architecture of existing work. This is due to the overly long token sequences and the interleaved text-image data format, which limit performance. To address these navigation challenges effectively, we propose Mobile-Agent-v2, a multi-agent architecture for mobile device operation assistance. The architecture comprises three agents: planning agent, decision agent, and reflection agent. The planning agent condenses lengthy, interleaved image-text history operations and screens summaries into a pure-text task progress, which is then passed on to the decision agent. This reduction in context length makes it easier for decision agent to navigate the task progress. To retain focus content, we design a memory unit that updates with task progress by decision agent. Additionally, to correct erroneous operations, the reflection agent observes the outcomes of each operation and handles any mistake accordingly. Experimental results indicate that Mobile-Agent-v2 achieves over a 30% improvement in task completion compared to the single-agent architecture of Mobile-Agent. The code is open-sourced at https: //github. com/X-PLUG/MobileAgent.

NeurIPS Conference 2024 Conference Paper

Training for Stable Explanation for Free

• Chao Chen
• Chenghua Guo
• Rufeng Chen
• Guixiang Ma
• Ming Zeng
• Xiangwen Liao
• Xi Zhang
• Sihong Xie

To foster trust in machine learning models, explanations must be faithful and stable for consistent insights. Existing relevant works rely on the $\ell_p$ distance for stability assessment, which diverges from human perception. Besides, existing adversarial training (AT) associated with intensive computations may lead to an arms race. To address these challenges, we introduce a novel metric to assess the stability of top-$k$ salient features. We introduce R2ET which trains for stable explanation by efficient and effective regularizer, and analyze R2ET by multi-objective optimization to prove numerical and statistical stability of explanations. Moreover, theoretical connections between R2ET and certified robustness justify R2ET's stability in all attacks. Extensive experiments across various data modalities and model architectures show that R2ET achieves superior stability against stealthy attacks, and generalizes effectively across different explanation methods. The code can be found at https: //github. com/ccha005/R2ET.

NeurIPS Conference 2024 Conference Paper

Trajectory Flow Matching with Applications to Clinical Time Series Modelling

• Xi Zhang
• Yuan Pu
• Yuki Kawamura
• Andrew Loza
• Yoshua Bengio
• Dennis L. Shung
• Alexander Tong

Modeling stochastic and irregularly sampled time series is a challenging problem found in a wide range of applications, especially in medicine. Neural stochastic differential equations (Neural SDEs) are an attractive modeling technique for this problem, which parameterize the drift and diffusion terms of an SDE with neural networks. However, current algorithms for training Neural SDEs require backpropagation through the SDE dynamics, greatly limiting their scalability and stability. To address this, we propose Trajectory Flow Matching (TFM), which trains a Neural SDE in a simulation-free manner, bypassing backpropagation through the dynamics. TFM leverages the flow matching technique from generative modeling to model time series. In this work we first establish necessary conditions for TFM to learn time series data. Next, we present a reparameterization trick which improves training stability. Finally, we adapt TFM to the clinical time series setting, demonstrating improved performance on four clinical time series datasets both in terms of absolute performance and uncertainty prediction, a crucial parameter in this setting.

AAAI Conference 2023 Conference Paper

Anytime User Engagement Prediction in Information Cascades for Arbitrary Observation Periods

• Akshay Aravamudan
• Xi Zhang
• Georgios C. Anagnostopoulos

Predicting user engagement -- whether a user will engage in a given information cascade -- is an important problem in the context of social media, as it is useful to online marketing and misinformation mitigation just to name a couple major applications. Based on split population multi-variate survival processes, we develop a discriminative approach that, unlike prior works, leads to a single model for predicting whether individual users of an information network will engage a given cascade for arbitrary forecast horizons and observation periods. Being probabilistic in nature, this model retains the interpretability of its generative counterpart and renders count prediction intervals in a disciplined manner. Our results indicate that our model is highly competitive, if not superior, to current approaches, when compared over varying observed cascade histories and forecast horizons.

EAAI Journal 2023 Journal Article

Graph features dynamic fusion learning driven by multi-head attention for large rotating machinery fault diagnosis with multi-sensor data

• Xin Zhang
• Xi Zhang
• Jie Liu
• Bo Wu
• Youmin Hu

Recently, rotating machinery fault diagnosis studies based on graph neural networks (GNN) have received some satisfactory achievements. But most of them are based on the analysis of the single sensor signals, which cannot capture the comprehensive fault information, especially aiming at large rotating machineries. A few research using GNN for multi-sensor fault diagnosis only fuse multi-source features in the construction of the input graph, and the fusion effect largely depends on the manual feature selection. Graph attention network (GAT), as an emerging GNN, can give trainable weights to vertices based on the self-attention mechanism to improve the effectiveness of feature learning. And it has not yet been used in the field of multi-sensor fault diagnosis. To fill this gap and utilize GAT’s advantages, this paper presents a multi-sensor multi-head GAT (MMHGAT) model for large rotating machinery fault diagnosis. With the input of several subgraphs, the designed MMHGAT model consisting of two graph attention layers (GAL), a feature fusion process and a Softmax classifier, can dynamically fuse and mine the high-level fault characteristics during the training process. By employing the experiment on the axial flow pump, the effectiveness and superiority of the proposed method are validated.

AAMAS Conference 2023 Conference Paper

M3: Modularization for Multi-task and Multi-agent Offline Pre-training

• Linghui Meng
• Jingqing Ruan
• Xuantang Xiong
• Xiyun Li
• Xi Zhang
• Dengpeng Xing
• Bo Xu

Learning a multi-task policy is crucial in multi-agent reinforcement learning (MARL). Recent work has focused on learning in the context of online multi-task reinforcement learning, where a policy is jointly trained from scratch, aiming to generalize well to few-shot or even zero-shot tasks. However, existing online methods require tremendous interactions and are therefore unsuitable for environments where interactions are expensive. In this work, we novelly introduce the modularization for multi-task and multi-agent offline pre-training (M3) to learn high-level transferable policy representations. We claim that the discrete policy representation is critical for multi-task offline learning and accordingly leverage contexts as a task prompt to enhance the adaptability of pre-trained models to various tasks. To disentangle multiple agents of variation under heterogeneous and non-stationary properties even though they receive the same task, we employ an agent-invariant VQ-VAE to identify each of the multiple agents. We encapsulate the pretrained model as part of an online MARL algorithm and fine-tune it * These authors contribute equally to this work. † Corresponding authors. Proc. of the 22nd International Conference on Autonomous Agents and Multiagent Systems (AAMAS 2023), A. Ricci, W. Yeoh, N. Agmon, B. An (eds.), May 29 – June 2, 2023, London, United Kingdom. © 2023 International Foundation for Autonomous Agents and Multiagent Systems (www. ifaamas. org). All rights reserved. to improve generalization. We also theoretically analyze the generalization error of our method. We test the proposed method on the challenging StarCraft Multi-Agent Challenge (SMAC) tasks, and empirical results show that it can achieve supreme performance in few-shot or even zero-shot settings across multiple tasks over state-of-the-art MARL methods.

AAAI Conference 2022 Conference Paper

DDGCN: Dual Dynamic Graph Convolutional Networks for Rumor Detection on Social Media

• Mengzhu Sun
• Xi Zhang
• Jiaqi Zheng
• Guixiang Ma

Detecting rumors on social media has become particular important due to the rapid dissemination and adverse impacts on our lives. Though a set of rumor detection models have exploited the message propagation structural or temporal information, they seldom model them altogether to enjoy the best of both worlds. Moreover, the dynamics of knowledge information associated with the comments are not involved, either. To this end, we propose a novel Dual-Dynamic Graph Convolutional Networks, termed as DDGCN, which can model the dynamics of messages in propagation as well as the dynamics of the background knowledge from Knowledge graphs in one unified framework. Specifically, two Graph Convolutional Networks are adopted to capture the above two types of structure information at different time stages, which are then combined with a temporal fusing unit. This allows for learning the dynamic event representations in a more fine-grained manner, and incrementally aggregating them to capture the cascading effect for better rumor detection. Extensive experiments on two public real-world datasets demonstrate that our proposal yields significant improvements compared to strong baselines and can detect rumors at early stages.

AAAI Conference 2022 Conference Paper

Explainable Survival Analysis with Convolution-Involved Vision Transformer

• Yifan Shen
• Li Liu
• Zhihao Tang
• Zongyi Chen
• Guixiang Ma
• Jiyan Dong
• Xi Zhang
• Lin Yang

Image-based survival prediction models can facilitate doctors in diagnosing and treating cancer patients. With the advance of digital pathology technologies, the big whole slide images (WSIs) provide increased resolution and more details for diagnosis. However, the gigabytesize or even terabyte-size WSIs would make most models computationally infeasible. To this end, instead of using the complete WSIs, most of the existing models only use a pre-selected subset of key patches or patch clusters as input, which might discard some important morphology information. In this work, we propose a novel survival analysis model to fully utilize the complete WSI information. We show that the use of a Vision Transformer (ViT) backbone, together with convolution operations involved in it, is an effective approach to improve the prediction performance. Additionally, we present a post-hoc explainable method to identify the most salient patches and distinct morphology features, making the model more faithful and the results easier to comprehend by human users. Evaluations on two large cancer datasets show that our proposed model is more effective and has better interpretability for survival prediction. We would make the code publicly available upon acceptance.

IJCAI Conference 2022 Conference Paper

MFAN: Multi-modal Feature-enhanced Attention Networks for Rumor Detection

• Jiaqi Zheng
• Xi Zhang
• Sanchuan Guo
• Quan Wang
• Wenyu Zang
• Yongdong Zhang

Rumor spreaders are increasingly taking advantage of multimedia content to attract and mislead news consumers on social media. Although recent multimedia rumor detection models have exploited both textual and visual features for classification, they do not integrate the social structure features simultaneously, which have shown promising performance for rumor identification. It is challenging to combine the heterogeneous multi-modal data in consideration of their complex relationships. In this work, we propose a novel Multi-modal Feature-enhanced Attention Networks (MFAN) for rumor detection, which makes the first attempt to integrate textual, visual, and social graph features in one unified framework. Specifically, it considers both the complement and alignment relationships between different modalities to achieve better fusion. Moreover, it takes into account the incomplete links in the social network data due to data collection constraints and proposes to infer hidden links to learn better social graph features. The experimental results show that MFAN can detect rumors effectively and outperform state-of-the-art methods.

JMLR Journal 2021 Journal Article

Collusion Detection and Ground Truth Inference in Crowdsourcing for Labeling Tasks

• Changyue Song
• Kaibo Liu
• Xi Zhang

Crowdsourcing has been a prompt and cost-effective way of obtaining labels in many machine learning applications. In the literature, a number of algorithms have been developed to infer the ground truth based on the collected labels. However, most existing studies assume workers to be independent and are vulnerable to worker collusion. This paper aims at detecting the collusive behaviors of workers in labeling tasks. Specifically, we consider collusion in a pairwise manner and propose a penalized pairwise profile likelihood method based on the adaptive LASSO penalty for collusion detection. Many models that describe the behavior of independent workers can be incorporated into our proposed framework as the baseline model. We further investigate the theoretical properties of the proposed method that guarantee the asymptotic performance. An algorithm based on expectation-maximization algorithm and coordinate descent is proposed to numerically maximize the penalized pairwise profile likelihood function for parameter estimation. To the best of our knowledge, this is the first statistical model that simultaneously detects collusion, learns workers’ capabilities, and infers the ground true labels. Numerical studies using synthetic and real data sets are also conducted to verify the performance of the method. [abs] [ pdf ][ bib ] &copy JMLR 2021. ( edit, beta )

NeurIPS Conference 2020 Conference Paper

On Numerosity of Deep Neural Networks

• Xi Zhang
• Xiaolin Wu

Recently, a provocative claim was published that number sense spontaneously emerges in a deep neural network trained merely for visual object recognition. This has, if true, far reaching significance to the fields of machine learning and cognitive science alike. In this paper, we prove the above claim to be unfortunately incorrect. The statistical analysis to support the claim is flawed in that the sample set used to identify number-aware neurons is too small, compared to the huge number of neurons in the object recognition network. By this flawed analysis one could mistakenly identify number-sensing neurons in any randomly initialized deep neural networks that are not trained at all. With the above critique we ask the question what if a deep convolutional neural network is carefully trained for numerosity? Our findings are mixed. Even after being trained with number-depicting images, the deep learning approach still has difficulties to acquire the abstract concept of numbers, a cognitive task that preschoolers perform with ease. But on the other hand, we do find some encouraging evidences suggesting that deep neural networks are more robust to distribution shift for small numbers than for large numbers.

IJCAI Conference 2020 Conference Paper

Rumor Detection on Social Media with Graph Structured Adversarial Learning

• Xiaoyu Yang
• Yuefei Lyu
• Tian Tian
• Yifei Liu
• Yudong Liu
• Xi Zhang

The wide spread of rumors on social media has caused tremendous effects in both the online and offline world. In addition to text information, recent detection methods began to exploit the graph structure in the propagation network. However, without a rigorous design, rumors may evade such graph models using various camouflage strategies by perturbing the structured data. Our focus in this work is to develop a robust graph-based detector to identify rumors on social media from an adversarial perspective. We first build a heterogeneous information network to model the rich information among users, posts, and user comments for detection. We then propose a graph adversarial learning framework, where the attacker tries to dynamically add intentional perturbations on the graph structure to fool the detector, while the detector would learn more distinctive structure features to resist such perturbations. In this way, our model would be enhanced in both robustness and generalization. Experiments on real-world datasets demonstrate that our model achieves better results than the state-of-the-art methods.

AAAI Conference 2019 Conference Paper

Cognitive Deficit of Deep Learning in Numerosity

• Xiaolin Wu
• Xi Zhang
• Xiao Shu

Subitizing, or the sense of small natural numbers, is an innate cognitive function of humans and primates; it responds to visual stimuli prior to the development of any symbolic skills, language or arithmetic. Given successes of deep learning (DL) in tasks of visual intelligence and given the primitivity of number sense, a tantalizing question is whether DL can comprehend numbers and perform subitizing. But somewhat disappointingly, extensive experiments of the type of cognitive psychology demonstrate that the examples-driven black box DL cannot see through superficial variations in visual representations and distill the abstract notion of natural number, a task that children perform with high accuracy and confidence. The failure is apparently due to the learning method not the CNN computational machinery itself. A recurrent neural network capable of subitizing does exist, which we construct by encoding a mechanism of mathematical morphology into the CNN convolutional kernels. Also, we investigate, using subitizing as a test bed, the ways to aid the black box DL by cognitive priors derived from human insight. Our findings are mixed and interesting, pointing to both cognitive deficit of pure DL, and some measured successes of boosting DL by predetermined cognitive implements. This case study of DL in cognitive computing is meaningful for visual numerosity represents a minimum level of human intelligence.

IJCAI Conference 2018 Conference Paper

CAGAN: Consistent Adversarial Training Enhanced GANs

• Yao Ni
• Dandan Song
• Xi Zhang
• Hao Wu
• Lejian Liao

Generative adversarial networks (GANs) have shown impressive results, however, the generator and the discriminator are optimized in finite parameter space which means their performance still need to be improved. In this paper, we propose a novel approach of adversarial training between one generator and an exponential number of critics which are sampled from the original discriminative neural network via dropout. As discrepancy between outputs of different sub-networks of a same sample can measure the consistency of these critics, we encourage the critics to be consistent to real samples and inconsistent to generated samples during training, while the generator is trained to generate consistent samples for different critics. Experimental results demonstrate that our method can obtain state-of-the-art Inception scores of 9. 17 and 10. 02 on supervised CIFAR-10 and unsupervised STL-10 image generation tasks, respectively, as well as achieve competitive semi-supervised classification results on several benchmarks. Importantly, we demonstrate that our method can maintain stability in training and alleviate mode collapse.

IJCAI Conference 2018 Conference Paper

Layered Optical Flow Estimation Using a Deep Neural Network with a Soft Mask

• Xi Zhang
• Di Ma
• Xu Ouyang
• Shanshan Jiang
• Lin Gan
• Gady Agam

Using a layered representation for motion estimation has the advantage of being able to cope with discontinuities and occlusions. In this paper, we learn to estimate optical flow by combining a layered motion representation with deep learning. Instead of pre-segmenting the image to layers, the proposed approach automatically generates a layered representation of optical flow using the proposed soft-mask module. The essential components of the soft-mask module are maxout and fuse operations, which enable a disjoint layered representation of optical flow and more accurate flow estimation. We show that by using masks the motion estimate results in a quadratic function of input features in the output layer. The proposed soft-mask module can be added to any existing optical flow estimation networks by replacing their flow output layer. In this work, we use FlowNet as the base network to which we add the soft-mask module. The resulting network is tested on three well-known benchmarks with both supervised and unsupervised flow estimation tasks. Evaluation results show that the proposed network achieve better results compared with the original FlowNet.

IJCAI Conference 2016 Conference Paper

Understanding Information Diffusion under Interactions

• Yuan Su
• Xi Zhang
• Philip S. Yu
• Wen Hua
• Xiaofang Zhou
• Binxing Fang

Information diffusion in online social networks has attracted substantial research effort. Although recent models begin to incorporate interactions among contagions, they still don't consider the comprehensive interactions involving users and contagions as a whole. Moreover, the interactions obtained in previous work are modeled as latent factors and thus are difficult to understand and interpret. In this paper, we investigate the contagion adoption behavior by incorporating various types of interactions into a coherent model, and propose a novel interaction-aware diffusion framework called IAD. IAD exploits the social network structures to distinguish user roles, and uses both structures and texts to categorize contagions. Experiments with large-scale Weibo dataset demonstrate that IAD outperforms the state-of-art baselines in terms of F1-score and accuracy, as well as the runtime for learning. In addition, the interactions obtained through learning reveal interesting findings, e. g. , food-related contagions have the strongest capability to suppress other contagions' propagation, while advertisement-related contagions have the weakest capability.

ICRA Conference 2015 Conference Paper

Tracking benchmark and evaluation for manipulation tasks

• Ankush Roy
• Xi Zhang
• Nina Wolleb
• Camilo Perez Quintero
• Martin Jägersand

In this paper we present a public dataset to evaluate trackers used for human and robot manipulation tasks. For these tasks both high DOF motion and high accuracy is needed. We describe in detail, both the process of recording the sequences and how ground truth data was generated for the videos. The videos are tagged with challenges that a tracker would face while tracking the object. As an initial example, we evaluate the performance of six published trackers [5], [11], [12], [13], [15], [6] and analyse their result. We describe a new evaluation metric to test sensitivity of trackers to speed. A total of 100 annotated and tagged sequences are reported. All the videos, ground truth data, original implementation of trackers and evaluation scripts are made publicly available on the website so others can extend the results on their trackers and evaluation.

AAAI Conference 2014 Conference Paper

Recommendation by Mining Multiple User Behaviors with Group Sparsity

• Ting Yuan
• Jian Cheng
• Xi Zhang
• Shuang Qiu
• Hanqing Lu

Recently, some recommendation methods try to improve the prediction results by integrating information from user’s multiple types of behaviors. How to model the dependence and independence between different behaviors is critical for them. In this paper, we propose a novel recommendation model, the Group-Sparse Matrix Factorization (GSMF), which factorizes the rating matrices for multiple behaviors into the user and item latent factor space with group sparsity regularization. It can (1) select out the different subsets of latent factors for different behaviors, addressing that users’ decisions on different behaviors are determined by different sets of factors; (2) model the dependence and independence between behaviors by learning the shared and private factors for multiple behaviors automatically; (3) allow the shared factors between different behaviors to be different, instead of all the behaviors sharing the same set of factors. Experiments on the real-world dataset demonstrate that our model can integrate users’ multiple types of behaviors into recommendation better, compared with other state-of-the-arts.

YNIMG Journal 2012 Journal Article

Altered spontaneous activity in Alzheimer's disease and mild cognitive impairment revealed by Regional Homogeneity

• Zengqiang Zhang
• Yong Liu
• Tianzi Jiang
• Bo Zhou
• Ningyu An
• Haitao Dai
• Pan Wang
• Yixuan Niu

Alzheimer's disease (AD), the most prevalent cause of dementia in the elderly, is characterized by progressive cognitive and intellectual deficits. Most patients with mild cognitive impairment (MCI) are thought to be in a very early stage of AD. Resting-state functional magnetic resonance imaging reflects spontaneous brain activities and/or the endogenous/background neurophysiological process of the human brain. Regional Homogeneity (ReHo) can provide a fast method for mapping regional activity across the whole brain. Little has been previously published about where or how spontaneous activity differs between MCI and AD, although many previous fMRI studies have shown that the activity pattern is altered in MCI/AD. In the present study, we first used the ReHo method to explore differences in regional spontaneous activities throughout the whole brain between normal controls (NC) and people with MCI and with AD. A one-way ANOVA was performed to determine the regions in which the ReHo differs between the three groups, and then a post hoc analysis was performed to evaluate differences in the pattern among the three groups. Finally a correlation analysis was done between the ReHo index of these regions and clinical variables in order to evaluate the relationship between ReHo and cognitive measures in the AD and MCI groups. An exploratory classification analysis also demonstrated that ReHo measures were able to correctly separate subjects in 71. 4% of the cases. Altered brain spontaneous activations were found in the medial prefrontal cortex, the bilateral posterior cingulate gyrus/precuneus and the left inferior parietal lobule (IPL) in both MCI and AD. In MCI, the ReHo index in the left IPL was higher than that of the NC, which could indicate the presence of a compensatory mechanism in MCI. More obviously, the correlation analysis indicated that the lower the memory and other cognitive abilities, the lower the ReHo in patients with MCI and AD. Combining our findings with the results in earlier studies, we propose that the spontaneous activity pattern in the resting state could potentially be used as a clinical marker for MCI/AD.