Hongyu Li

EAAI Journal 2026 Journal Article

A new closed-loop structure belief rule base for complex systems

• Zongjun Zhang
• Wei He
• Ning Ma
• Hongyu Li
• Guohui Zhou

Reasonable modeling is an important aspect of complex system management and decision-making. Belief rule base (BRB) is a powerful tool for modeling complex systems. In the face of the complexity of the system and the potential risk of failure, the decision support capability of the model is the key to ensure the effectiveness of modeling. However, BRB models for decision support face two problems: the difficulty of balancing the interpretability and accuracy of the process, and the difficulty of tracing the causes of the results. Therefore, a new closed-loop structure BRB (CLBRB) for complex systems is proposed, where the closed-loop structure consists of an interpretable process and traceable results. First, this paper proposes a method to quantify the interpretability of the BRB model. Based on this, three interpretability enhancement strategies and a new assessment metric are proposed to achieve an interpretable adaptive balance between accuracy and interpretability to enhance the interpretability of the process. Second, a new reverse causal inference (RCI-R) model based on reverse causal inference rule matrix (RCRM) and a corresponding rule modeling and inference scheme are proposed to track and parse the influencing factors behind the results and accomplish closed-loop decision-making. The effectiveness and superiority of the CLBRB model are verified by taking the health state assessment of the aerospace relay and lithium-ion battery as examples.

AAAI Conference 2026 Conference Paper

AerialVLA: A Vision-Language-Action Model for Aerial Navigation with Online Dialogue

• Jinyu Chen
• Hongyu Li
• Zongheng Tang
• Xiaoduo Li
• Wenjun Wu
• Si Liu

Visual Dialogue Navigation (VDN) aims to enable agents to reach target locations through dialogue with humans. The integration of VDN into Unmanned Aerial Vehicle (UAV) systems enhances human-machine interaction by enabling intuitive, hands-free operation, thereby unlocking vast applications. However, existing VDN models for UAVs can only perform navigation based on dialogue history, lacking proactive interaction capabilities to correct trajectories. Moreover, their sequential observation history recording mechanism struggles to accurately localize landmarks observed in the historical context, leading to ineffective utilization of referential information in new user instructions.To address these, we present AerialVLA, an end-to-end UAV navigation framework integrating dialogue comprehension, action decision-making, and navigational question generation. AerialVLA comprises three core components: i) we propose the Progress-Driven Navigation-Query Alternation mechanism to determine optimal questioning timing through navigation progress estimation autonomously. ii) To effectively model long-horizon history observation sequences, we develop the History Spatial-Temporal Fusion module that extracts discriminative spatial-temporal representations from historical observations. iii) Furthermore, to overcome data scarcity in training, we devise the Online Task-Driven Augmentation strategy that enhances learning through action-conditioned data augmentation. Experimental results demonstrate that AerialVLA achieves state-of-the-art navigation performance while exhibiting effective dialogue capabilities.Moreover, to better evaluate the agent's proactive dialogue and navigation abilities, our evaluation benchmark, named UAV Navigation with Online Dialogue (UNOD), incorporates an online dialogue interaction module. The UNOD assesses UAV agents' real-time questioning capabilities by leveraging an Air Commander Large Language Model to simulate human-UAV interactions during testing.

AAAI Conference 2026 Conference Paper

Shadows in the Code: Exploring the Risks and Defenses of LLM-based Multi-Agent Software Development Systems

• Xiaoqing Wang
• Keman Huang
• Bin Liang
• Hongyu Li
• Xiaoyong Du

The rapid advancement of Large Language Model (LLM)-driven multi-agent systems has significantly streamlined software developing tasks, enabling users with little technical expertise to develop executable applications. While these systems democratize software creation through natural language requirements, they introduce significant security risks that remain largely unexplored. We identify two risky scenarios: Malicious User with Benign Agents (MU-BA) and Benign User with Malicious Agents (BU-MA). We introduce the Implicit Malicious Behavior Injection Attack (IMBIA), demonstrating how multi-agent systems can be manipulated to generate software with concealed malicious capabilities beneath seemingly benign applications, and propose Adv-IMBIA as a defense mechanism. Evaluations across ChatDev, MetaGPT, and AgentVerse frameworks reveal varying vulnerability patterns, with IMBIA achieving attack success rates of 93%, 45%, and 71% in MU-BA scenarios, and 71%, 84%, and 45% in BU-MA scenarios. Our defense mechanism reduced attack success rates significantly, particularly in the MU-BA scenario. Further analysis reveals that compromised agents in the coding and testing phases pose significantly greater security risks, while also identifying critical agents that require protection against malicious user exploitation. Our findings highlight the urgent need for robust security measures in multi-agent software development systems and provide practical guidelines for implementing targeted, resource-efficient defensive strategies.

AAAI Conference 2026 Conference Paper

Time-Frequency Token Advantage Clipping for Training Efficient Large Reasoning Model

• Rong Bao
• Bo Wang
• Xiao Wang
• Hongyu Li
• Rui Zheng
• Leszek Rutkowski
• Qi Zhang
• Liang Ding

Long Chain-of-Thought (CoT) reasoning enhances large reasoning models' performance but suffers from severe inefficiencies, as models often overthink simple problems or underthink complex ones. Current sequence-level optimizations, like length penalties, are too coarse-grained to distinguish core logic from verbose language, precluding the necessary token-level control for efficient reasoning CoT. To overcome these limitations, we introduce Time-Frequency token Advantage Clipping (TFAC), a novel training framework designed to build efficient large reasoning models via token-level interventions. Specifically, TFAC functions along two dimensions: 1) The Frequency Dimension: It discourages inefficient loops and encourages deeper exploration by dynamically reducing the advantage scores of high-entropy tokens that are repeatedly generated within a single reasoning path. 2) The Time Dimension: It reduces excessive overthinking of the system by establishing a historical baseline for the occurrence count of each critical token in previously successful trajectories, and clipping the advantages of tokens that exceed this baseline during training. Crucially, to preserve the model's exploratory capabilities on novel problems, this suppression mechanism is automatically disabled when no historical record of success is available. Experiments conducted on the Deepseek-Distill-32B and Qwen3-8B models show that TFAC outperforms leading baseline methods, improving performance by 2.3 and 3.1 percentage points, respectively, while simultaneously reducing inference costs by 35% and 28% in scenarios where correct answers are generated. These results validate the significant efficacy of TFAC in training large reasoning models that are both powerful and highly efficient.

ICLR Conference 2025 Conference Paper

Point Cluster: A Compact Message Unit for Communication-Efficient Collaborative Perception

• Zihan Ding
• Jiahui Fu 0003
• Si Liu 0001
• Hongyu Li
• Siheng Chen
• Hongsheng Li 0001
• Shifeng Zhang
• Xu Zhou

The objective of the collaborative perception task is to enhance the individual agent's perception capability through message communication among neighboring agents. A central challenge lies in optimizing the inherent trade-off between perception ability and communication cost. To tackle this bottleneck issue, we argue that a good message unit should encapsulate both semantic and structural information in a sparse format, a feature not present in prior approaches. In this paper, we innovatively propose a compact message unit, namely point cluster, whose core idea is to represent potential objects efficiently with explicitly decoupled low-level structure information and high-level semantic information. Building upon this new message unit, we propose a comprehensive framework CPPC for communication-efficient collaborative perception. The core principle of CPPC is twofold: first, through strategical point sampling, structure information can be well preserved with a few key points, which can significantly reduce communication cost; second, the sequence format of point clusters enables efficient message aggregation by set matching and merging, thereby eliminating unnecessary computation generated when aligning squared BEV maps, especially for long-range collaboration. To handle time latency and pose errors encountered in real-world scenarios, we also carefully design parameter-free solutions that can adapt to different noisy levels without finetuning. Experiments on two widely recognized collaborative perception benchmarks showcase the superior performance of our method compared to the previous state-of-the-art approaches.

NeurIPS Conference 2025 Conference Paper

Self-Evolving Pseudo-Rehearsal for Catastrophic Forgetting with Task Similarity in LLMs

• Jun Wang
• Liang Ding
• Shuai Wang
• Hongyu Li
• Yong Luo
• Huangxuan Zhao
• Han Hu
• Bo Du

Continual learning for large language models (LLMs) demands a precise balance between $\textbf{plasticity}$ - the ability to absorb new tasks - and $\textbf{stability}$ - the preservation of previously learned knowledge. Conventional rehearsal methods, which replay stored examples, are limited by long-term data inaccessibility; earlier pseudo-rehearsal methods require additional generation modules, while self-synthesis approaches often generate samples that poorly align with real tasks, suffer from unstable outputs, and ignore task relationships. We present $\textbf{\textit{Self-Evolving Pseudo-Rehearsal for Catastrophic Forgetting with Task Similarity}}(\textbf{SERS})$, a lightweight framework that 1) decouples pseudo-input synthesis from label creation, using semantic masking and template guidance to produce diverse, task-relevant prompts without extra modules; 2) applies label self-evolution, blending base-model priors with fine-tuned outputs to prevent over-specialization; and 3) introduces a dynamic regularizer driven by the Wasserstein distance between task distributions, automatically relaxing or strengthening constraints in proportion to task similarity. Experiments across diverse tasks on different LLMs show that our SERS reduces forgetting by over 2\% points against strong pseudo-rehearsal baselines, by ensuring efficient data utilization and wisely transferring knowledge. The code will be released at https: //github. com/JerryWangJun/LLM_CL_SERS/.

AAAI Conference 2025 Conference Paper

Unleashing the Temporal-Spatial Reasoning Capacity of GPT for Training-Free Audio and Language Referenced Video Object Segmentation

• Shaofei Huang
• Rui Ling
• Hongyu Li
• Tianrui Hui
• Zongheng Tang
• Xiaoming Wei
• Jizhong Han
• Si Liu

In this paper, we propose an Audio-Language-Referenced SAM 2 (AL-Ref-SAM 2) pipeline to explore the training-free paradigm for audio and language-referenced video object segmentation, namely AVS and RVOS tasks. The intuitive solution leverages GroundingDINO to identify the target object from a single frame and SAM 2 to segment the identified object throughout the video, which is less robust to spatiotemporal variations due to a lack of video context exploration. Thus, in our AL-Ref-SAM 2 pipeline, we propose a novel GPT-assisted Pivot Selection (GPT-PS) module to instruct GPT-4 to perform two-step temporal-spatial reasoning for sequentially selecting pivot frames and pivot boxes, thereby providing SAM 2 with a high-quality initial object prompt. Within GPT-PS, two task-specific Chain-of-Thought prompts are designed to unleash GPT’s temporal-spatial reasoning capacity by guiding GPT to make selections based on a comprehensive understanding of video and reference information. Furthermore, we propose a Language-Binded Reference Unification (LBRU) module to convert audio signals into language-formatted references, thereby unifying the formats of AVS and RVOS tasks in the same pipeline. Extensive experiments show that our training-free AL-Ref-SAM 2 pipeline achieves performances comparable to or even better than fully-supervised fine-tuning methods.

AAAI Conference 2023 Conference Paper

Dual Mutual Information Constraints for Discriminative Clustering

• Hongyu Li
• Lefei Zhang
• Kehua Su

Deep clustering is a fundamental task in machine learning and data mining that aims at learning clustering-oriented feature representations. In previous studies, most of deep clustering methods follow the idea of self-supervised representation learning by maximizing the consistency of all similar instance pairs while ignoring the effect of feature redundancy on clustering performance. In this paper, to address the above issue, we design a dual mutual information constrained clustering method named DMICC which is based on deep contrastive clustering architecture, in which the dual mutual information constraints are particularly employed with solid theoretical guarantees and experimental validations. Specifically, at the feature level, we reduce the redundancy among features by minimizing the mutual information across all the dimensionalities to encourage the neural network to extract more discriminative features. At the instance level, we maximize the mutual information of the similar instance pairs to obtain more unbiased and robust representations. The dual mutual information constraints happen simultaneously and thus complement each other to jointly optimize better features that are suitable for the clustering task. We also prove that our adopted mutual information constraints are superior in feature extraction, and the proposed dual mutual information constraints are clearly bounded and thus solvable. Extensive experiments on five benchmark datasets show that our proposed approach outperforms most other clustering algorithms. The code is available at https://github.com/Li-Hyn/DMICC.

YNIMG Journal 2020 Journal Article

Acceleration of three-dimensional diffusion magnetic resonance imaging using a kernel low-rank compressed sensing method

• Chaoyi Zhang
• Tanzil Mahmud Arefin
• Ukash Nakarmi
• Choong Heon Lee
• Hongyu Li
• Dong Liang
• Jiangyang Zhang
• Leslie Ying

Diffusion Magnetic Resonance Imaging (dMRI) has shown great potential in probing tissue microstructure and structural connectivity in the brain but is often limited by the lengthy scan time needed to sample the diffusion profile by acquiring multiple diffusion weighted images (DWIs). Although parallel imaging technique has improved the speed of dMRI acquisition, attaining high resolution three dimensional (3D) dMRI on preclinical MRI systems remained still time consuming. In this paper, kernel principal component analysis, a machine learning approach, was employed to estimate the correlation among DWIs. We demonstrated the feasibility of such correlation estimation from low-resolution training DWIs and used the correlation as a constraint to reconstruct high-resolution DWIs from highly under-sampled k-space data, which significantly reduced the scan time. Using full k-space 3D dMRI data of post-mortem mouse brains, we retrospectively compared the performance of the so-called kernel low rank (KLR) method with a conventional compressed sensing (CS) method in terms of image quality and ability to resolve complex fiber orientations and connectivity. The results demonstrated that the KLR-CS method outperformed the conventional CS method for acceleration factors up to 8 and was likely to enhance our ability to investigate brain microstructure and connectivity using high-resolution 3D dMRI.

IJCAI Conference 2009 Conference Paper

• Chao Zhang
• Hongyu Li
• Qiyong Guo
• Jinyuan Jia
• I-Fan Shen

This paper proposes a novel framework for image retrieval. The retrieval is treated as searching for an ordered cycle in an image database. The optimal cycle can be found by minimizing the geometric manifold entropy of images. The minimization is solved by the proposed method, fast active tabu search. Experimental results demonstrate the framework for image retrieval is feasible and quite promising.

IJCAI Conference 2005 Conference Paper

Supervised Local Tangent Space Alignment for Classification

• Hongyu Li
• Wenbin Chen
• I-Fan

Supervised local tangent space alignment (SLTSA) is an extension of local tangent space alignment (LTSA) to supervised feature extraction. Two algorithmic improvements are made upon LTSA for classification. First a simple technique is proposed to map new data to the embedded low-dimensional space and make LTSA suitable in a changing, dynamic environment. Then SLTSA is introduced to deal with data sets containing multiple classes with class membership information.