arXiv subjects

Peng Yu

Publications and source records attributed to Peng Yu.

At least 19 recordsLinked to original sources

arXiv2026

DMRL: Document-Mediated Reinforcement Learning for Skill Optimization in Advertising Recommendation

Advertising recommendation requires continuously tuning complex system parameters while balancing commercial returns and user experience. Recent work has introduced large language models (LLMs) with skill documents to assist this labor-intensive process, but skill optimization remains largely prompt-driven, lacking a principled mechanism to attribute rewards to specific document edits. To address this limitation, we propose Document-Mediated Reinforcement Learning (DMRL), a skill self-evolution framework that models skill document optimization as a sequence of structured editing actions. In DMRL, an upper-level agent performs controlled document edits, while a frozen lower-level task agent evaluates their effects through A/B testing. To address credit assignment and long-term outcomes, we introduce two key components: (1) Dual-Relative Policy Optimization (DRPO), a post-training policy optimization method for robust and risk-aware advantage estimation; and (2) Long-term Reward Predictor (LRP), which estimates long-term outcomes by modeling population heterogeneity with disentangled representation learning and cross-attention transfer. DMRL was deployed on a large-scale short-video ads platform and extensive empirical evaluation shows that DMRL outperforms state-of-the-art baselines across key advertising metrics

cs.LG↗

arXiv2026

Fair ASR: Re-Evaluating Black-Box Jailbreaks under Shared Target-Call Budgets

Reliable jailbreak evaluation is essential for assessing LLM safety, but most existing studies rely solely on attack success rate (ASR) without accounting for its dependence on attack budgets, resulting in unfair comparisons across methods. Existing compute-aware evaluations reduce heterogeneous resources into FLOPs, which is difficult to estimate for black-box models and fails to capture resource-specific constraints. To provide a comparable evaluation basis, we introduce Fair-ASR, an evaluation protocol for black-box jailbreak attacks under shared target-call budgets B, using target calls as a directly observable and method-agnostic comparison axis while tracking attacker calls separately for efficiency analysis. We re-evaluate 11 representative attacks under the Fair-ASR protocol and find that attack rankings change substantially across target-call budgets, simple stochastic perturbations and hand-crafted templates remain highly competitive under equal target access, and no evaluated LLM-driven method is efficient in both target and attacker calls. Motivated by this efficiency gap, we introduce ReCode, a compositional budget-efficient attack that combines desensitization rewriting with two effective low-cost primitives identified by Fair-ASR. Under a budget of 20 target calls, ReCode achieves 85% ASR on GPT-5 while requiring only 7.19 attacker calls per request on average, showing strong efficiency in both target and attacker calls.

cs.CR↗

arXiv2026

JailbreakSkill: Scaling Automated Red-Teaming with Reusable and Ever-Evolving Skills

Automated red-teaming has produced a growing collection of attack strategies, yet they typically remain scattered across prompts and workflows, making them difficult to systematically integrate, reuse, and improve at scale. We introduce \textsc{JailbreakSkill}, a skill-centric framework for scaling automated red-teaming through reusable and continuously evolving attack capabilities. \textsc{JailbreakSkill} packages existing attack strategies into modular, agent-ready skills that can be directly reused and adaptively selected across tasks and target models. Beyond reuse, it closes the loop between attacking and learning: attack experience is used to diagnose, refine, combine, and discover new skills, which are added back to an ever-growing skill library. This evolution lifts macro-average ASR by 17.5 percentage points on AdvBench and 13.4 points on HarmBench, including a 48.6-point gain against GPT-5.4 on AdvBench, while yielding novel attack strategies such as reframing a direct request as an unfinished document-completion task. Several evolved skills also generalize to unseen prompts and target models without further adaptation. Our code is available at https://github.com/BattleWen/JailbreakSkill.

cs.AI↗

arXiv2026

Spatiotemporal Graph Transformer for Traffic Intelligence in Edge Computing

Accurate traffic forecasting is essential for proactive resource management in edge computing, where service demand evolves dynamically across both space and time. In practical cellular edge systems, traffic exhibits strong spatial correlations among neighboring service regions and long-range temporal dependencies driven by user mobility and application behavior. Existing recurrent forecasting approaches can capture short-term dynamics but often struggle to model long-horizon traffic evolution under non-stationary conditions. To address this challenge, we propose a spatiotemporal graph Transformer framework that jointly models spatial interactions and temporal dependencies for traffic forecasting in edge computing. The framework employs graph neural networks to capture spatial correlations among service regions and leverages Transformer-based self-attention to learn long-range temporal patterns from historical traffic observations. By decoupling spatial representation learning from temporal reasoning, the proposed approach provides an effective mechanism for large-scale spatiotemporal traffic modeling. Extensive experiments on a real-world cellular network dataset demonstrate that the proposed graph Transformer consistently outperforms recurrent graph-based baselines, including GCN-RNN, GCN-LSTM, and GCN-GRU models, across multiple forecasting horizons. The resulting forecasts enable more effective proactive resource provisioning and reduce overload risk compared with reactive management strategies. These results highlight the potential of graph-enhanced attention mechanisms for building intelligent and adaptive edge computing systems.

cs.LG↗

arXiv2026

An Early Warning of Emerging Biosecurity Risks in Frontier LLMs

Frontier large language models (LLMs) are increasingly integrated into scientific workflows, yet their growing biological capabilities may outpace current safeguards. To assess the biological risks of frontier models, we develop Intern-BioBreaker, a specialized bio-red-teaming model, together with an integrated computational-to-physical framework that couples model-level stress testing with wet-lab validation. Within this framework, Intern-BioBreaker generates targeted jailbreak prompts to test whether aligned models can be induced to provide operational guidance for safety-sensitive biological tasks or produce sequence-level outputs with potentially harmful properties. Selected sequence outputs are then carried forward for DNA synthesis, host expression, and orthogonal protein verification to assess whether model-generated designs can yield the intended biological products. Our evaluation reveals a concerning gap between text-level safeguards and the risks posed by capable scientific models: (i) Intern-BioBreaker outperforms baseline attack models and reveals widespread bio-risk jailbreak vulnerabilities across both open-weight and proprietary frontier LLMs, with several targets reaching near-saturated or 100% task-level attack success rate (ASR); (ii) in sequence-level case studies, GPT-5.5 can be induced to generate modified viral candidate sequences with pathogenic potential; the corresponding translated proteins may exhibit even stronger receptor-binding affinity and thus enhanced infection potential; and (iii) end-to-end verification shows that selected model-generated biological designs are not merely textual artifacts, but can be physically realized under controlled experimental settings. These findings underscore the need for stronger biological red-teaming, nucleic acid synthesis screening, and safety mechanisms that keep pace with model capabilities.

cs.CL↗

arXiv2026

Arachne: Orchestrating Cascades for Efficient Text-to-Video Model Training

The rising demand for AI-generated videos is fueled by advances in large-scale Text-to-Video (T2V) models, trained on extensive datasets of video clips spanning diverse resolutions and durations. To address this data heterogeneity, current training methods often use a bucketing strategy that groups samples into discrete buckets for efficiency. However, this approach struggles to scale with compute and data volumes under static parallelism schemes, such as data and sequence parallelism, leading to significant workload imbalances and hardware under-utilization. In this paper, we present Arachne, a novel training framework for efficient T2V model training at scale. Arachne decomposes the training process into fine-grained computational units, called \textit{cascades}, orchestrating their distributed execution and synchronization across the cluster through coordinated spatial and temporal optimization. Our comprehensive evaluation demonstrates that Arachne reduces iteration time by up to 65\% over leading frameworks, exhibiting a positive scaling trend where its performance advantages amplify as training scale grows.

cs.DC↗

arXiv2026

Not All Turns Matter: Credit Assignment for Multi-Turn Jailbreaking

Deploying LLMs in multi-turn dialogues facilitates jailbreak attacks that distribute harmful intent across seemingly benign turns. Recent training-based multi-turn jailbreak methods learn long-horizon attack strategies from interaction feedback, but often rely on coarse trajectory-level outcome signals that broadcast uniformly to every turn. However, we find that turn-level contributions in multi-turn jailbreaking are non-uniform, phase-dependent, and target-specific. Such coarse outcome supervision induces a credit assignment problem, leading to over-rewarding redundant turns in successful trajectories and under-crediting useful intermediate turns in failed ones. To address this, we propose TRACE, a turn-aware credit assignment framework for reinforcement learning (RL)-based multi-turn jailbreaking. For successful trajectories, TRACE estimates turn-level contributions via leave-one-turn-out semantic masking; for failed ones, TRACE assigns penalties based on prompt harmfulness and semantic relevance, with an additional local refusal-aware penalty. Furthermore, we reuse the attack-side credit signal for multi-turn defense alignment. Extensive experiments on open-source and closed-source targets show that TRACE achieves strong overall performance in effectiveness, transferability, and efficiency, yielding about a 25% relative improvement in attack success rate over the strongest RL baseline while also improving the safety-utility balance when reused for defense alignment.

cs.AI↗

arXiv2026

Quadrature-TreeSHAP: Depth-Independent TreeSHAP and Shapley Interactions

Shapley values are a standard tool for explaining predictions of tree ensembles, with Path-Dependent SHAP being the most widely used variant. Despite substantial progress, existing methods still exhibit trade-offs between depth-dependent runtime, numerical stability, and support for higher-order interactions. To address these challenges, we introduce Quadrature-TreeSHAP, a quadrature-based reformulation of Path-Dependent TreeSHAP that is numerically stable, naturally extends to any-order Shapley interaction values and is practically insensitive to tree depth. Our implementation supports both CPU and GPU and is integrated into XGBoost. Our method is based on a weighted-Banzhaf interaction polynomial, which expresses Banzhaf interaction values as expectations under a feature participation probability $p$. Shapley values and any-order interaction values are then recovered by integrating these polynomials over $p$ from 0 to 1. We evaluate these integrals using Gauss-Legendre quadrature, and show that, in practice, only 8 fixed quadrature points are sufficient to reach machine precision. In fact, Quadrature-TreeSHAP with 8 fixed points achieves greater numerical stability than TreeSHAP. This fixed-point formulation removes depth dependence from the inner computation and enables efficient SIMD execution. We confirm these advantages empirically. On 12 XGBoost benchmarks, Quadrature-TreeSHAP computes Shapley values 1.06x-10.59x faster than TreeSHAP on CPU and 1.84x-6.95x faster than GPUTreeSHAP on GPU. Shapley pairwise interactions are 3.80x-58.11x faster on CPU, with higher-order interactions achieving speedups of up to 1200x compared to TreeSHAP-IQ.

cs.LG↗

arXiv2026

STEM: Structure-Tracing Evidence Mining for Knowledge Graphs-Driven Retrieval-Augmented Generation

Knowledge Graph-based Question Answering (KGQA) plays a pivotal role in complex reasoning tasks but remains constrained by two persistent challenges: the structural heterogeneity of Knowledge Graphs(KGs) often leads to semantic mismatch during retrieval, while existing reasoning path retrieval methods lack a global structural perspective. To address these issues, we propose Structure-Tracing Evidence Mining (STEM), a novel framework that reframes multi-hop reasoning as a schema-guided graph search task. First, we design a Semantic-to-Structural Projection pipeline that leverages KG structural priors to decompose queries into atomic relational assertions and construct an adaptive query schema graph. Subsequently, we execute globally-aware node anchoring and subgraph retrieval to obtain the final evidence reasoning graph from KG. To more effectively integrate global structural information during the graph construction process, we design a Triple-Dependent GNN (Triple-GNN) to generate a Global Guidance Subgraph (Guidance Graph) that guides the construction. STEM significantly improves both the accuracy and evidence completeness of multi-hop reasoning graph retrieval, and achieves State-of-the-Art performance on multiple multi-hop benchmarks.

cs.CL↗

arXiv2026

First Law for Nonsingular Black Holes in 2D Dilaton Gravity

A central issue in the thermodynamics of nonsingular black holes is the apparent violation of the first law. In this work, we use 2D dilaton gravity as a simple theoretical setting to study this issue. We systematically construct a broad class of nonsingular black hole solutions with metric function $A(x)=f(x)+c$, through a procedure that is considerably simpler than in higher-dimensional theories. Using the Iyer-Wald covariant phase space formalism, we derive the correct energy formula and establish a consistent first law for this entire class of solutions. The apparent violation of the first law in a previous work arises because the energy used therein is not the Hamiltonian conjugate to the fixed time-translation generator adopted. Under the boundary conditions and normalization convention specified in the present work, the Iyer-Wald Hamiltonian energy is $E=-\frac{c}{2}$, and the first law is restored. Moreover, the energy formula agrees with the Casimir function in 2D dilaton gravity, thus confirming its interpretation as the physical black hole energy. Our results clarify the correct first law for 2D nonsingular black holes and may provide insights into the first law of nonsingular black holes in higher dimensions.

gr-qc↗

arXiv2026

Shape-Interpretable Visual Self-Modeling Enables Geometry-Aware Continuum Robot Control

Continuum robots possess high flexibility and redundancy, making them well suited for safe interaction in complex environments, yet their continuous deformation and nonlinear dynamics pose fundamental challenges to perception, modeling, and control. Existing vision-based control approaches often rely on end-to-end learning, achieving shape regulation without explicit awareness of robot geometry or its interaction with the environment. Here, we introduce a shape-interpretable visual self-modeling framework for continuum robots that enables geometry-aware control. Robot shapes are encoded from multi-view planar images using a Bezier-curve representation, transforming visual observations into a compact and physically meaningful shape space that uniquely characterizes the robot's three-dimensional configuration. Based on this representation, neural ordinary differential equations are employed to self-model both shape and end-effector dynamics directly from data, enabling hybrid shape-position control without analytical models or dense body markers. The explicit geometric structure of the learned shape space allows the robot to reason about its body and surroundings, supporting environment-aware behaviors such as obstacle avoidance and self-motion while maintaining end-effector objectives. Experiments on a cable-driven continuum robot demonstrate accurate shape-position regulation and tracking, with shape errors within 1.56% of image resolution and end-effector errors within 2% of robot length, as well as robust performance in constrained environments. By elevating visual shape representations from two-dimensional observations to an interpretable three-dimensional self-model, this work establishes a principled alternative to vision-based end-to-end control and advances autonomous, geometry-aware manipulation for continuum robots.

cs.RO↗

arXiv2026

Tele-Omni: a Unified Multimodal Framework for Video Generation and Editing

Recent advances in diffusion-based video generation have substantially improved visual fidelity and temporal coherence. However, most existing approaches remain task-specific and rely primarily on textual instructions, limiting their ability to handle multimodal inputs, contextual references, and diverse video generation and editing scenarios within a unified framework. Moreover, many video editing methods depend on carefully engineered pipelines tailored to individual operations, which hinders scalability and composability. In this paper, we propose Tele-Omni, a unified multimodal framework for video generation and editing that follows multimodal instructions, including text, images, and reference videos, within a single model. Tele-Omni leverages pretrained multimodal large language models to parse heterogeneous instructions and infer structured generation or editing intents, while diffusion-based generators perform high-quality video synthesis conditioned on these structured signals. To enable joint training across heterogeneous video tasks, we introduce a task-aware data processing pipeline that unifies multimodal inputs into a structured instruction format while preserving task-specific constraints. Tele-Omni supports a wide range of video-centric tasks, including text-to-video generation, image-to-video generation, first-last-frame video generation, in-context video generation, and in-context video editing. By decoupling instruction parsing from video synthesis and combining it with task-aware data design, Tele-Omni achieves flexible multimodal control while maintaining strong temporal coherence and visual consistency. Experimental results demonstrate that Tele-Omni achieves competitive performance across multiple tasks.

cs.CV↗

arXiv2026

TeleBoost: A Systematic Alignment Framework for High-Fidelity, Controllable, and Robust Video Generation

Post-training is the decisive step for converting a pretrained video generator into a production-oriented model that is instruction-following, controllable, and robust over long temporal horizons. This report presents a systematical post-training framework that organizes supervised policy shaping, reward-driven reinforcement learning, and preference-based refinement into a single stability-constrained optimization stack. The framework is designed around practical video-generation constraints, including high rollout cost, temporally compounding failure modes, and feedback that is heterogeneous, uncertain, and often weakly discriminative. By treating optimization as a staged, diagnostic-driven process rather than a collection of isolated tricks, the report summarizes a cohesive recipe for improving perceptual fidelity, temporal coherence, and prompt adherence while preserving the controllability established at initialization. The resulting framework provides a clear blueprint for building scalable post-training pipelines that remain stable, extensible, and effective in real-world deployment settings.

cs.CV↗

arXiv2026

Unusually high phonon thermal conductivity in the Weyl semimetal TaP: A comparative study with TaAs

In many metals, thermal transport is often dominated by electrons, although the lattice contribution can remain appreciable depending on the material. Here, through rigorous first-principles calculations, we uncover a phonon-dominated thermal transport regime in the Weyl semimetals TaAs and TaP. Remarkably, despite its metallic character, TaP exhibits an exceptionally high phonon thermal conductivity ($\kappa_{\rm ph}$) of 162 Wm$^{-1}$K$^{-1}$ at room temperature, surpassing its electronic counterpart by nearly an order of magnitude. This anomalously high $\kappa_{\rm ph}$ arises from the combined effects of Weyl electronic structure, acoustic phonon bunching, and a large low-high frequency phonon gap, which together suppress phonon-electron and three-phonon scattering. The linearly dispersing bands near the Fermi level yield a low electronic density of states, limiting both electrical conductivity and electronic thermal transport in these compounds. By further surveying a broad range of topological semimetals, we show that the prominence of phonon thermal transport is a universal characteristic of this material class. Our work provides deeper insight into thermal transport mechanisms in topological semimetals and broadens the scope for discovering metals with high thermal conductivity.

cond-mat.mtrl-sci↗

arXiv2025

Enhanced superconducting diode effect in hybrid Josephson junctions

The superconducting diode effect (SDE) has recently been observed in various systems, sparking interest in novel superconducting devices and offering a new platform to probe intrinsic material properties. Josephson junctions with strong Rashba spin-orbit coupling have exhibited nonreciprocal critical currents under applied magnetic fields. In this work, we investigate the SDE in Josephson junctions incorporating periodic hole arrays patterned into the superconducting leads on InAs heterostructures with epitaxial aluminum. We observe an enhanced diode effect when a top gate depletes the 2DEG in the region of the hole arrays, while preserving the overall supercurrent. Theoretical analysis shows that the physics behind this phenomenon is the increased difference of transparency between different bands in the junction. These results highlight a new pathway for engineering and controlling nonreciprocal superconducting transport in hybrid systems.

cond-mat.mes-hall↗

arXiv2025

Binary Split Categorical feature with Mean Absolute Error Criteria in CART

In the context of the Classification and Regression Trees (CART) algorithm, the efficient splitting of categorical features using standard criteria like GINI and Entropy is well-established. However, using the Mean Absolute Error (MAE) criterion for categorical features has traditionally relied on various numerical encoding methods. This paper demonstrates that unsupervised numerical encoding methods are not viable for the MAE criteria. Furthermore, we present a novel and efficient splitting algorithm that addresses the challenges of handling categorical features with the MAE criterion. Our findings underscore the limitations of existing approaches and offer a promising solution to enhance the handling of categorical data in CART algorithms.

cs.LG↗

arXiv2025

Anisotropic Hot Carrier Relaxation and Coherent Phonon Dynamics in Type-II Weyl Semimetal TaIrTe4

The unique energy band and crystal structure of the layered type-II Weyl semimetal TaIrTe4 hold great promise for high-performance broadband anisotropic optoelectronic devices. Therefore, gaining an in-depth understanding of the interactions between internal microscopic particles is of vital importance. Here, we employ a two-color pump-probe system to reveal the anisotropic electron-phonon coupling (EPC) and coherent phonon dynamics in bulk TaIrTe4. The carrier relaxation exhibits a four-exponential decay process, with strong dependence on polarization of probe pulse, indicating that EPC strength is closely related to the crystal axes (a/b-axes). In addition, we observe three coherent phonon modes in bulk TaIrTe4: 38.5 GHz, 0.44 THz and 1.29 THz. Their oscillation amplitudes and dephasing times also show anisotropic responses to the probe polarization. We also investigate the in-plane cross-directional thermal conductivity coefficient of TaIrTe4 by beam-offset frequency-domain thermal reflection (FDTR). The thermal conductivity coefficient along the a-axis and b-axis directions are ka=14.4 W/mK and kb=3.8 W/mK, respectively. This represents a significant in-plane anisotropy. Our work not only reveals the key role of anisotropic EPC in controlling the thermal and optical properties of TaIrTe4, but also provides insights into designing polarization-sensitive optoelectronic devices based on topological semimetals.

physics.optics↗

arXiv2025

PC-NCLaws: Physics-Embedded Conditional Neural Constitutive Laws for Elastoplastic Materials

While data-driven methods offer significant promise for modeling complex materials, they often face challenges in generalizing across diverse physical scenarios and maintaining physical consistency. To address these limitations, we propose a generalizable framework called Physics-Embedded Conditional Neural Constitutive Laws for Elastoplastic Materials, which combines the partial differential equations with neural networks. Specifically, the model employs two separate neural networks to model elastic and plastic constitutive laws. Simultaneously, the model incorporates physical parameters as conditional inputs and is trained on comprehensive datasets encompassing multiple scenarios with varying physical parameters, thereby enabling generalization across different properties without requiring retraining for each individual case. Furthermore, the differentiable architecture of our model, combined with its explicit parameter inputs, enables the inverse estimation of physical parameters from observed motion sequences. This capability extends our framework to objects with unknown or unmeasured properties. Experimental results demonstrate state-of-the-art performance in motion reconstruction, robust long-term prediction, geometry generalization, and precise parameters estimation for elastoplastic materials, highlighting its versatility as a unified simulator and inverse analysis tool.

cs.GR↗