arXiv subjects

Hongzong Li

Publications and source records attributed to Hongzong Li.

14 recordsLinked to original sources

arXiv2026

Where World Models Break: Natural-Input Failure Discovery

World models predict action-conditioned futures and serve as critical internal simulators for downstream planning and control. However, catastrophic prediction failures of world models could dangerously propagate through the control pipeline, as subsequent agent or model training and decision-making depend heavily on the continuous environment evolution forecasted by these world models. Existing evaluations overlook this systemic risk: by aggregating average errors over benign generations from general queries, they fail to stress-test the model against catastrophic collapses under rare or unobserved condition-action combinations. To bridge this gap, we formalize the natural-input failure discovery problem: under a finite query budget, finding environment-valid conditions and action prefixes that induce severe prediction risk, verifying whether these failures reproduce on fresh seeds, and testing their persistence under nearby valid edits. Discovering such critical failures is computationally challenging, as valid condition-action combinations explode exponentially, rendering exhaustive search or standard sampling infeasible given the high cost of noisy rollouts. To tackle this, we propose BasinLens, which exploits the underlying structure of valid inputs, where each coordinate possesses environment-defined semantic types and admissible domains, by pairing uncertainty-guided global search with typed local replacements. Across diverse benchmarks and world-model families, BasinLens exposes reproducible and locally persistent failure modes that conventional evaluations fail to reveal, showing that average-case benchmarks can mask important vulnerabilities in world-model-driven control.

cs.AI↗

arXiv2026

Efficient Online Lexicographic Generalized Low-Rank Matrix Bandits

This paper studies generalized low-rank matrix bandits with multiple prioritized objectives. At each round, the learner selects a matrix-valued arm and observes a vector-valued reward, whose components correspond to multiple objectives with different priority levels. Each objective is governed by an objective-specific generalized low-rank matrix model, and the learner evaluates arms according to a lexicographic preference order, prioritizing higher-level objectives before lower-level ones. We propose \textsc{Lexi-LowGLM}, an efficient online algorithm that first estimates objective-specific low-rank subspaces and then performs lexicographic learning in the reduced feature spaces. Unlike existing single-objective algorithms that repeatedly solve a batch generalized linear estimator using all historical observations, \textsc{Lexi-LowGLM} updates each objective-specific estimator via an online Newton step, reducing the estimator-update complexity over $T$ rounds from $O(T^2)$ to $O(T)$. We establish a regret bound of $\widetilde O\left(W_i^{\rm lex}\sqrt{m}\,(d_1+d_2)r\sqrt{T}\right)$ for each objective $i\in[m]$, where $r$ is an upper bound on the ranks of the objective-specific parameter matrices and $W_i^{\rm lex}$ characterizes the lexicographic trade-off effect. This bound depends on the effective low-rank dimension $(d_1+d_2)r$ rather than the ambient dimension $d_1d_2$. Numerical experiments further validate the effectiveness and computational efficiency of the proposed method.

cs.LG↗

arXiv2026

How Vulnerable Are Edge LLMs?

Large language models (LLMs) are increasingly deployed on edge devices under strict computation and quantization constraints, yet their security implications remain unclear. We study query-based knowledge extraction from quantized edge-deployed LLMs under realistic query budgets and show that, although quantization introduces noise, it does not remove the underlying semantic knowledge, allowing substantial behavioral recovery through carefully designed queries. To systematically analyze this risk, we propose \textbf{CLIQ} (\textbf{Cl}ustered \textbf{I}nstruction \textbf{Q}uerying), a structured query construction framework that improves semantic coverage while reducing redundancy. Experiments on quantized Qwen models (INT8/INT4) demonstrate that CLIQ consistently outperforms original queries across BERTScore, BLEU, and ROUGE, enabling more efficient extraction under limited budgets. These results indicate that quantization alone does not provide effective protection against query-based extraction, highlighting a previously underexplored security risk in edge-deployed LLMs.

cs.CR↗

arXiv2026

How Much Information Can a Vision Token Hold? A Scaling Law for Recognition Limits in VLMs

Recent vision-centric approaches have made significant strides in long-context modeling. Represented by DeepSeek-OCR, these models encode rendered text into continuous vision tokens, achieving high compression rates without sacrificing recognition precision. However, viewing the vision encoder as a lossy channel with finite representational capacity raises a fundamental question: what is the information upper bound of visual tokens? To investigate this limit, we conduct controlled stress tests by progressively increasing the information quantity (character count) within an image. We observe a distinct phase-transition phenomenon characterized by three regimes: a near-perfect Stable Phase, an Instability Phase marked by increased error variance, and a total Collapse Phase. We analyze the mechanical origins of these transitions and identify key factors. Furthermore, we formulate a probabilistic scaling law that unifies average vision token load and visual density into a latent difficulty metric. Extensive experiments across various Vision-Language Models demonstrate the universality of this scaling law, providing critical empirical guidance for optimizing the efficiency-accuracy trade-off in visual context compression.

cs.LG↗

arXiv2025

Balancing Efficiency and Fairness: An Iterative Exchange Framework for Multi-UAV Cooperative Path Planning

Multi-UAV cooperative path planning (MUCPP) is a fundamental problem in multi-agent systems, aiming to generate collision-free trajectories for a team of unmanned aerial vehicles (UAVs) to complete distributed tasks efficiently. A key challenge lies in achieving both efficiency, by minimizing total mission cost, and fairness, by balancing the workload among UAVs to avoid overburdening individual agents. This paper presents a novel Iterative Exchange Framework for MUCPP, balancing efficiency and fairness through iterative task exchanges and path refinements. The proposed framework formulates a composite objective that combines the total mission distance and the makespan, and iteratively improves the solution via local exchanges under feasibility and safety constraints. For each UAV, collision-free trajectories are generated using A* search over a terrain-aware configuration space. Comprehensive experiments on multiple terrain datasets demonstrate that the proposed method consistently achieves superior trade-offs between total distance and makespan compared to existing baselines.

cs.RO↗

arXiv2025

Particle Dynamics for Latent-Variable Energy-Based Models

Latent-variable energy-based models (LVEBMs) assign a single normalized energy to joint pairs of observed data and latent variables, offering expressive generative modeling while capturing hidden structure. We recast maximum-likelihood training as a saddle problem over distributions on the latent and joint manifolds and view the inner updates as coupled Wasserstein gradient flows. The resulting algorithm alternates overdamped Langevin updates for a joint negative pool and for conditional latent particles with stochastic parameter ascent, requiring no discriminator or auxiliary networks. We prove existence and convergence under standard smoothness and dissipativity assumptions, with decay rates in KL divergence and Wasserstein-2 distance. The saddle-point view further yields an ELBO strictly tighter than bounds obtained with restricted amortized posteriors. Our method is evaluated on numerical approximations of physical systems and performs competitively against comparable approaches.

cs.LG↗

arXiv2025

Adversary-Free Counterfactual Prediction via Information-Regularized Representations

We study counterfactual prediction under assignment bias and propose a mathematically grounded, information-theoretic approach that removes treatment-covariate dependence without adversarial training. Starting from a bound that links the counterfactual-factual risk gap to mutual information, we learn a stochastic representation Z that is predictive of outcomes while minimizing I(Z; T). We derive a tractable variational objective that upper-bounds the information term and couples it with a supervised decoder, yielding a stable, provably motivated training criterion. The framework extends naturally to dynamic settings by applying the information penalty to sequential representations at each decision time. We evaluate the method on controlled numerical simulations and a real-world clinical dataset, comparing against recent state-of-the-art balancing, reweighting, and adversarial baselines. Across metrics of likelihood, counterfactual error, and policy evaluation, our approach performs favorably while avoiding the training instabilities and tuning burden of adversarial schemes.

cs.LG↗

arXiv2025

Physicochemically Informed Dual-Conditioned Generative Model of T-Cell Receptor Variable Regions for Cellular Therapy

Physicochemically informed biological sequence generation has the potential to accelerate computer-aided cellular therapy, yet current models fail to \emph{jointly} ensure novelty, diversity, and biophysical plausibility when designing variable regions of T-cell receptors (TCRs). We present \textbf{PhysicoGPTCR}, a large generative protein Transformer that is \emph{dual-conditioned} on peptide and HLA context and trained to autoregressively synthesise TCR sequences while embedding residue-level physicochemical descriptors. The model is optimised on curated TCR--peptide--HLA triples with a maximum-likelihood objective and compared against ANN, GPTCR, LSTM, and VAE baselines. Across multiple neoantigen benchmarks, PhysicoGPTCR substantially improves edit-distance, similarity, and longest-common-subsequence scores, while populating a broader region of sequence space. Blind in-silico docking and structural modelling further reveal a higher proportion of binding-competent clones than the strongest baseline, validating the benefit of explicit context conditioning and physicochemical awareness. Experimental results demonstrate that dual-conditioned, physics-grounded generative modelling enables end-to-end design of functional TCR candidates, reducing the discovery timeline from months to minutes without sacrificing wet-lab verifiability.

cs.CE↗

arXiv2025

ABConformer: Physics-inspired Sliding Attention for Antibody-Antigen Interface Prediction

Accurate prediction of antibody-antigen (Ab-Ag) interfaces is critical for vaccine design, immunodiagnostics, and therapeutic antibody development. However, achieving reliable predictions from sequences alone remains a challenge. In this paper, we present ABCONFORMER, a model based on the Conformer backbone that captures both local and global features of a biosequence. To accurately capture Ab-Ag interactions, we introduced the physics-inspired sliding attention, enabling residue-level contact recovery without relying on three-dimensional structural data. ABConformer can accurately predict paratopes and epitopes given the antibody and antigen sequence, and predict pan-epitopes on the antigen without antibody information. In comparison experiments, ABCONFORMER achieves state-of-the-art performance on a recent SARS-CoV-2 Ab-Ag dataset, and surpasses widely used sequence-based methods for antibody-agnostic epitope prediction. Ablation studies further quantify the contribution of each component, demonstrating that, compared to conventional cross-attention, sliding attention significantly enhances the precision of epitope prediction. To facilitate reproducibility, we will release the code under an open-source license upon acceptance.

cs.LG↗

arXiv2025

CAME-AB: Cross-Modality Attention with Mixture-of-Experts for Antibody Binding Site Prediction

Antibody binding site prediction plays a pivotal role in computational immunology and therapeutic antibody design. Existing sequence or structure methods rely on single-view features and fail to identify antibody-specific binding sites on the antigens. In this paper, we propose \textbf{CAME-AB}, a novel Cross-modality Attention framework with a Mixture-of-Experts (MoE) backbone for robust antibody binding site prediction. CAME-AB integrates five biologically grounded modalities, including raw amino acid encodings, BLOSUM substitution profiles, pretrained language model embeddings, structure-aware features, and GCN-refined biochemical graphs, into a unified multimodal representation. To enhance adaptive cross-modal reasoning, we propose an \emph{adaptive modality fusion} module that learns to dynamically weight each modality based on its global relevance and input-specific contribution. A Transformer encoder combined with an MoE module further promotes feature specialization and capacity expansion. We additionally incorporate a supervised contrastive learning objective to explicitly shape the latent space geometry, encouraging intra-class compactness and inter-class separability. To improve optimization stability and generalization, we apply stochastic weight averaging during training. Extensive experiments on benchmark antibody-antigen datasets demonstrate that CAME-AB consistently outperforms strong baselines on multiple metrics, including Precision, Recall, F1-score, AUC-ROC, and MCC. Ablation studies further validate the effectiveness of each architectural component and the benefit of multimodal feature integration. The model implementation details and the codes are available on https://anonymous.4open.science/r/CAME-AB-C525

cs.LG↗

arXiv2025

TopoBind: Multi-Modal Prediction of Antibody-Antigen Binding Free Energy via Sequence Embeddings and Structural Topology

Predicting the binding free energy between antibodies and antigens is a key challenge in structure-aware biomolecular modeling, with direct implications for antibody design. Most existing methods either rely solely on sequence embeddings or struggle to capture complex structural relationships, thus limiting predictive performance. In this work, we present a novel framework that integrates sequence-based representations from pre-trained protein language models (ESM-2) with a set of topological features. Specifically, we extract contact map metrics reflecting residue-level connectivity, interface geometry descriptors characterizing cross-chain interactions, distance map statistics quantifying spatial organization, and persistent homology invariants that systematically capture the emergence and persistence of multi-scale topological structures - such as connected components, cycles, and cavities - within individual proteins and across the antibody-antigen interface. By leveraging a cross-attention mechanism to fuse these diverse modalities, our model effectively encodes both global and local structural organization, thereby substantially enhancing the prediction of binding free energy. Extensive experiments demonstrate that our model consistently outperforms sequence-only and conventional structural models, achieving state-of-the-art accuracy in binding free energy prediction.

q-bio.BM↗

arXiv2025

BConformeR: A Conformer Based on Mutual Sampling for Unified Prediction of Continuous and Discontinuous Antibody Binding Sites

Accurate prediction of antibody-binding sites (epitopes) on antigens is crucial for vaccine design, immunodiagnostics, therapeutic antibody development, antibody engineering, research into autoimmune and allergic diseases, and advancing our understanding of immune responses. Despite in silico methods that have been proposed to predict both linear (continuous) and conformational (discontinuous) epitopes, they consistently underperform in predicting conformational epitopes. In this work, we propose Conformer-based models trained separately on AlphaFold-predicted structures and experimentally determined structures, leveraging convolutional neural networks (CNNs) to extract local features and Transformers to capture long-range dependencies within antigen sequences. Ablation studies demonstrate that CNN enhances the prediction of linear epitopes, and the Transformer module improves the prediction of conformational epitopes. Experimental results show that our model outperforms existing baselines in terms of MCC, ROC-AUC, PR-AUC, and F1 scores on both linear and conformational epitopes.

q-bio.BM↗

arXiv2025

Catch Causal Signals from Edges for Label Imbalance in Graph Classification

Despite significant advancements in causal research on graphs and its application to cracking label imbalance, the role of edge features in detecting the causal effects within graphs has been largely overlooked, leaving existing methods with untapped potential for further performance gains. In this paper, we enhance the causal attention mechanism through effectively leveraging edge information to disentangle the causal subgraph from the original graph, as well as further utilizing edge features to reshape graph representations. Capturing more comprehensive causal signals, our design leads to improved performance on graph classification tasks with label imbalance issues. We evaluate our approach on real-word datasets PTC, Tox21, and ogbg-molhiv, observing improvements over baselines. Overall, we highlight the importance of edge features in graph causal detection and provide a promising direction for addressing label imbalance challenges in graph-level tasks. The model implementation details and the codes are available on https://github.com/fengrui-z/ECAL

cs.LG↗

arXiv2025

Remodeling Peptide-MHC-TCR Triad Binding as Sequence Fusion for Immunogenicity Prediction

The complex nature of tripartite peptide-MHC-TCR interactions is a critical yet underexplored area in immunogenicity prediction. Traditional studies on TCR-antigen binding have not fully addressed the complex dependencies in triad binding. In this paper, we propose new modeling approaches for these tripartite interactions, utilizing sequence information from MHCs, peptides, and TCRs. Our methods adhere to native sequence forms and align with biological processes to enhance prediction accuracy. By incorporating representation learning techniques, we introduce a fusion mechanism to integrate the three sequences effectively. Empirical experiments show that our models outperform traditional methods, achieving a 2.8 to 13.3 percent improvement in prediction accuracy across existing benchmarks. We further validate our approach with extensive ablation studies, demonstrating the effectiveness of the proposed model components. The model implementation, code, and supplementary materials, including a manuscript with colored hyperlinks and a technical appendix for digital viewing, will be open-sourced upon publication.

q-bio.BM↗