arXiv subjects

Jack Yang

Publications and source records attributed to Jack Yang.

10 recordsLinked to original sources

arXiv2026

SHIELD: An Auto-Healing Agentic Defense Framework for LLM Resource Exhaustion Attacks

Sponge attacks increasingly threaten LLM systems by inducing excessive computation and DoS. Existing defenses either rely on statistical filters that fail on semantically meaningful attacks or use static LLM-based detectors that struggle to adapt as attack strategies evolve. We introduce SHIELD, a multi-agent, auto-healing defense framework centered on a three-stage Defense Agent that integrates semantic similarity retrieval, pattern matching, and LLM-based reasoning. Two auxiliary agents, a Knowledge Updating Agent and a Prompt Optimization Agent, form a closed self-healing loop, when an attack bypasses detection, the system updates an evolving knowledgebase, and refines defense instructions. Extensive experiments show that SHIELD consistently outperforms perplexity-based and standalone LLM defenses, achieving high F1 scores across both non-semantic and semantic sponge attacks, demonstrating the effectiveness of agentic self-healing against evolving resource-exhaustion threats.

cs.CR↗

arXiv2025

Prompt-Induced Over-Generation as Denial-of-Service: A Black-Box Attack-Side Benchmark

Large Language Models (LLMs) can be driven into over-generation, emitting thousands of tokens before producing an end-of-sequence (EOS) token. This degrades answer quality, inflates latency and cost, and can be weaponized as a denial-of-service (DoS) attack. Recent work has begun to study DoS-style prompt attacks, but typically focuses on a single attack algorithm or assumes white-box access, without an attack-side benchmark that compares prompt-based attackers in a black-box, query-only regime with a known tokenizer. We introduce such a benchmark and study two prompt-only attackers. The first is an Evolutionary Over-Generation Prompt Search (EOGen) that searches the token space for prefixes that suppress EOS and induce long continuations. The second is a goal-conditioned reinforcement learning attacker (RL-GOAL) that trains a network to generate prefixes conditioned on a target length. To characterize behavior, we introduce Over-Generation Factor (OGF): the ratio of produced tokens to a model's context window, along with stall and latency summaries. EOGen discovers short-prefix attacks that raise Phi-3 to OGF = 1.39 +/- 1.14 (Success@>=2: 25.2%); RL-GOAL nearly doubles severity to OGF = 2.70 +/- 1.43 (Success@>=2: 64.3%) and drives budget-hit non-termination in 46% of trials.

cs.CR↗

arXiv2025

MACE Foundation Models for Lattice Dynamics: A Benchmark Study on Double Halide Perovskites

Recent developments in materials informatics and artificial intelligence has led to the emergence of foundational energy models for material chemistry, as represented by the suite of MACE-based foundation models, bringing a significant breakthrough in universal potentials for inorganic solids. As to all method developments in computational materials science, performance benchmarking against existing high-level data with focusing on specific applications, is critically needed to understand the limitations in the models, thus facilitating the ongoing improvements in the model development process, and occasionally, leading to significant conceptual leaps in materials theory. Here, using our own published DFT (Density Functional Theory) database of room-temperature dynamic stability and vibrational anharmonicity for $\sim2000$ cubic halide double perovskites, we benchmarked the performances of four different variants of the MACE foundation models for screening the dynamic stabilities of inorganic solids. Our analysis shows that, as anticipated, the model accuracy improves with more training data. The dynamic stabilities of weakly anharmonic materials (as predicted by DFT) are more accurately reproduced by the foundation model, than those highly anharmonic and dynamically unstable ones. The predominant source of error in predicting the dynamic stability arises predominantly from the amplification of errors in atomic forces when predicting the harmonic phonon properties through the computation of the Hessian matrix, less so is the contribution from possible differences in the range of the configurational spaces that are sampled by DFT and the foundation model in molecular dynamics. We hope that our present findings will stimulate future works towards more physics-inspired approaches in assessing the accuracy of foundation models for atomistic modelling.

cond-mat.mtrl-sci↗

arXiv2025

V-GameGym: Visual Game Generation for Code Large Language Models

Code large language models have demonstrated remarkable capabilities in programming tasks, yet current benchmarks primarily focus on single modality rather than visual game development. Most existing code-related benchmarks evaluate syntax correctness and execution accuracy, overlooking critical game-specific metrics such as playability, visual aesthetics, and user engagement that are essential for real-world deployment. To address the gap between current LLM capabilities in algorithmic problem-solving and competitive programming versus the comprehensive requirements of practical game development, we present V-GameGym, a comprehensive benchmark comprising 2,219 high-quality samples across 100 thematic clusters derived from real-world repositories, adopting a novel clustering-based curation methodology to ensure both diversity and structural completeness. Further, we introduce a multimodal evaluation framework with an automated LLM-driven pipeline for visual code synthesis using complete UI sandbox environments. Our extensive analysis reveals that V-GameGym effectively bridges the gap between code generation accuracy and practical game development workflows, providing quantifiable quality metrics for visual programming and interactive element generation.

cs.SE↗

arXiv2025

Spatio-Temporal Joint Density Driven Learning for Skeleton-Based Action Recognition

Traditional approaches in unsupervised or self supervised learning for skeleton-based action classification have concentrated predominantly on the dynamic aspects of skeletal sequences. Yet, the intricate interaction between the moving and static elements of the skeleton presents a rarely tapped discriminative potential for action classification. This paper introduces a novel measurement, referred to as spatial-temporal joint density (STJD), to quantify such interaction. Tracking the evolution of this density throughout an action can effectively identify a subset of discriminative moving and/or static joints termed "prime joints" to steer self-supervised learning. A new contrastive learning strategy named STJD-CL is proposed to align the representation of a skeleton sequence with that of its prime joints while simultaneously contrasting the representations of prime and nonprime joints. In addition, a method called STJD-MP is developed by integrating it with a reconstruction-based framework for more effective learning. Experimental evaluations on the NTU RGB+D 60, NTU RGB+D 120, and PKUMMD datasets in various downstream tasks demonstrate that the proposed STJD-CL and STJD-MP improved performance, particularly by 3.5 and 3.6 percentage points over the state-of-the-art contrastive methods on the NTU RGB+D 120 dataset using X-sub and X-set evaluations, respectively.

cs.CV↗

arXiv2025

Causal Discovery in Symmetric Dynamic Systems with Convergent Cross Mapping

This paper systematically discusses how the inherent properties of chaotic attractors influence the results of discovering causality from time series using convergent cross mapping, particularly how convergent cross mapping misleads bidirectional causality as unidirectional when the chaotic attractor exhibits symmetry. We propose a novel method based on the k-means clustering method to address the challenges when the chaotic attractor exhibits two-fold rotation symmetry. This method is demonstrated to recover the symmetry of the latent chaotic attractor and discover the correct causality between time series without introducing information from other variables. We validate the accuracy of this method using time series derived from low-dimension and high-dimensional chaotic symmetric attractors for which convergent cross mapping may conclude erroneous results.

math.DS↗

arXiv2024

FullStack Bench: Evaluating LLMs as Full Stack Coders

As the capabilities of code large language models (LLMs) continue to expand, their applications across diverse code intelligence domains are rapidly increasing. However, most existing datasets only evaluate limited application domains. To address this gap, we have developed a comprehensive code evaluation dataset FullStack Bench focusing on full-stack programming, which encompasses a wide range of application domains (e.g., basic programming, data analysis, software engineering, mathematics, and machine learning). Besides, to assess multilingual programming capabilities, in FullStack Bench, we design real-world instructions and corresponding unit test cases from 16 widely-used programming languages to reflect real-world usage scenarios rather than simple translations. Moreover, we also release an effective code sandbox execution tool (i.e., SandboxFusion) supporting various programming languages and packages to evaluate the performance of our FullStack Bench efficiently. Comprehensive experimental results on our FullStack Bench demonstrate the necessity and effectiveness of our FullStack Bench and SandboxFusion.

cs.AI↗

arXiv2024

Joint Temporal Pooling for Improving Skeleton-based Action Recognition

In skeleton-based human action recognition, temporal pooling is a critical step for capturing spatiotemporal relationship of joint dynamics. Conventional pooling methods overlook the preservation of motion information and treat each frame equally. However, in an action sequence, only a few segments of frames carry discriminative information related to the action. This paper presents a novel Joint Motion Adaptive Temporal Pooling (JMAP) method for improving skeleton-based action recognition. Two variants of JMAP, frame-wise pooling and joint-wise pooling, are introduced. The efficacy of JMAP has been validated through experiments on the popular NTU RGB+D 120 and PKU-MMD datasets.

cs.CV↗

arXiv2023

Strengthening our grip on food security by encoding physics into AI

Climate change will jeopardize food security. Food security involves the robustness of the global agri-food system. This agri-food system is intricately connected to systems centering around health, economy, social-cultural diversity, and global political stability. A systematic way to determine acceptable interventions in the global agri-food systems involves analyses at different spatial and temporal scales. Such multi-scale analyses are common within physics. Unfortunately, physics alone is not sufficient. Machine learning techniques may aid. We focus on neural networks (NN) into which physics-based information is encoded (PeNN) and apply it to a sub-problem within the agri-food system. We show that the mean squared error of the PeNN is always smaller than that of the NNs, in the order of a factor of thousand. Furthermore, the PeNNs capture extra and interpolation very well, contrary to the NNs. It is shown that PeNNs need a much smaller data set size than the NNs to achieve a similar mse. Our results suggest that the incorporation of physics into neural networks architectures yields promise for addressing food security.

physics.soc-ph↗

arXiv2007

Effect of Internal Viscosity on Brownian Dynamics of DNA Molecules in Shear Flow

The results of Brownian dynamics simulations of a single DNA molecule in shear flow are presented taking into account the effect of internal viscosity. The dissipative mechanism of internal viscosity is proved necessary in the research of DNA dynamics. A stochastic model is derived on the basis of the balance equation for forces acting on the chain. The Euler method is applied to the solution of the model. The extensions of DNA molecules for different Weissenberg numbers are analyzed. Comparison with the experimental results available in the literature is carried out to estimate the contribution of the effect of internal viscosity.

q-bio.BM↗