arXiv subjects

Zhipeng Chen

Publications and source records attributed to Zhipeng Chen.

At least 19 recordsLinked to original sources

arXiv2026

SphereVAE: Hyperspherical Latent Autoencoders for Robust Autoregressive Speech Representation Modeling

With the rapid development of speech generation technology, discrete codec representations have been widely used because they provide a stable prediction paradigm. In expressive speech generation, however, the quantization bottleneck of discrete codecs results in information gaps in fine-grained prosody, timbre, pronunciation, and frame-to-frame continuity. Continuous representations (e.g., VAE latents), by eliminating this constraint, have emerged as a more effective alternative for autoregressive modeling. Yet when continuous representations are used as autoregressive prediction targets, prediction errors can accumulate along the generation chain, causing latent drift and degrading long-form stability. To mitigate this problem, we propose SphereVAE, which constrains the VAE latent space to the unit hypersphere. SphereVAE defines a Power Spherical posterior on the hypersphere and regularizes the latent distribution toward a uniform prior, so that information is encoded mainly by directional variation, providing a bounded geometric target for autoregressive prediction and reducing the risk of norm drift. SphereVAE underperforms the standard VAE on reconstruction metrics due to reduced latent freedom. However, when integrated into VoxCPM for zero-shot TTS and long-text generation, it yields lower content error rates with comparable speaker similarity, and shows more stable long-range speaker consistency. These results indicate that an appropriate latent geometric constraint can effectively mitigate autoregressive error accumulation and drift in speech generation.

eess.AS↗

arXiv2026

Quaternary-Squeeze Quantum Identity Authentication: Direction-Scaling Security via Single-Mode Squeezed Light

Quantum identity authentication (QIA) has emerged as a crucial technology for secure communication systems, particularly in the burgeoning era of quantum communications. This paper proposes a novel QIA protocol based on non-classical characteristics of squeezed light fields. By exploiting quantum noise reduction properties of quadrature squeezed coherent states, the protocol fundamentally thwarts eavesdropping attempts by Heisenberg-limited uncertainty constraints. The fidelity parameter for decoy states is utilized to detect spoofing attacks, and the dynamic key update mechanism fundamentally eliminates vulnerabilities caused by key reuse. Security information ratio analysis shows that the protocol is able to resist Gaussian-cloner attacks and detect eavesdropping. Moreover, the security threshold can be further enhanced with higher squeezing, allowing tunable protection levels adaptable to different threat scenarios. Compared with binary-squeezed protocols, our proposed four-direction (quaternary-dimensional) squeezing halves the eavesdropper's guessing probability and enlarges the fidelity gap by 29% tightening the discrimination threshold without relying on extra hardware, thus facilitating practical implementation.

quant-ph↗

arXiv2026

Effects of external magnetic field ripple on FRC equilibrium

The two-dimensional equilibrium of Field-Reversed Configuration (FRC) plasma in presence of an external ripple magnetic field is computed to show the emergence of multiple magnetic axes for hollow equilibrium current profiles. An increase in ripple amplitude reduces the hollowness threshold required for the development of multiple magnetic axes. For an intermediate range of the ripple axial period, the formation of multiple magnetic axes becomes the most likely. The ripple's radial extension and the curvature of the axial field are the critical factors underlying the non-monotonic effect of ripple's axial period. When the axial period increases from the lower range, the ripple's radial extension gradually grows and enhances the chance of forming multiple magnetic axes. Once the ripple's radial extension covers the entire radial domain, further increasing the ripple axial period decreases the ripple field curvature, which becomes the dominant factor for lowering the hollowness threshold for the formation of multiple magnetic axes.

physics.plasm-ph↗

arXiv2026

MotionEnhancer: Leveraging Video Diffusion for Motion-Enhanced Vision-Language Models

The new era has witnessed a remarkable capability to extend Vision-Language Models (VLMs) for tackling tasks of video understanding. While current VLMs excel at event- or story-level understanding, their ability to capture fine-grained motion details remains limited, primarily due to their focus on high-level static semantic structures and macro-event logic. In contrast, Video Diffusion Models (VDMs) are adept at modeling dynamic motion patterns, benefiting from large-scale video data and the intrinsic requirement of temporal generation. In this paper, we introduce MotionEnhancer, a novel approach that leverages motion priors distilled from a powerful video diffusion model as auxiliary supervision to enhance the motion understanding capability of a VLM via attention alignment. MotionEnhancer comprises two simple parameter-free modules, Motion-sensitive Head Selection (MHS) and Motion-salient Text Token Identification (MTTI), to directly extract and optimize motion-related attentions from the VDM in a computation-only manner. MotionEnhancer provides a scalable solution for motion understanding without additional training parameters, modifications to existing architectures, or tool calling. Extensive experiments demonstrate that MotionEnhancer can achieve consistent improvements over state-of-the-art VLMs on two motion-level video understanding benchmarks, especially on motion-related metrics.

cs.CV↗

arXiv2026

Low-rank Optimization Trajectories Modeling for LLM RLVR Acceleration

Recently, scaling reinforcement learning with verifiable rewards (RLVR) for large language models (LLMs) has emerged as an effective training paradigm for significantly improving model capabilities, which requires guiding the model to perform extensive exploration and learning, leading to substantial computational overhead and becoming a key challenge. To reduce the number of training steps, Prior work performs linear extrapolation of model parameters. However, the dynamics of model parameter updates during RLVR training remain insufficiently understood. To further investigate the evolution of LLMs during RLVR training, we conduct empirical experiments and find that the rank-1 subspace of the model does not evolve linearly, and its dominance over the original parameters is further amplified during LoRA training. Based on the above insights, we propose the \textbf{N}onlinear \textbf{Ext}rapolation of low-rank trajectories (\textbf{NExt}), a novel framework that models and extrapolates low-rank parameter trajectories in a nonlinear manner. Concretely, we first train the model using LoRA and extract the rank-1 subspace of parameter differences at multiple training steps, which is then used for the subsequent nonlinear extrapolation. Afterward, we utilized the extracted rank-1 subspace to train a predictor, which can model the trajectory of parameter updates during RLVR, and then perform the predict-extend process to extrapolate model parameters, achieving the acceleration of RLVR. To further study and understand NExt, we conduct comprehensive experiments that demonstrate the effectiveness and robustness of the method. Our method reduces computational overhead by approximately 37.5\% while remaining compatible with a wide range of RLVR algorithms and tasks. We release our code in https://github.com/RUCAIBox/NExt.

cs.LG↗

arXiv2026

Towards Effective Experiential Learning: Dual Guidance for Utilization and Internalization

Recently, reinforcement learning~(RL) has become an important approach for improving the capabilities of large language models~(LLMs). In particular, reinforcement learning from verifiable rewards~(RLVR) has emerged as a promising paradigm for reasoning tasks. However, existing RL-based training still remains only a rough approximation to human learning. Human learners leverage both external and internal experience to guide exploration and gradually internalize useful trajectories into stable knowledge. Motivated by this gap, we ask: how can LLMs better utilize and internalize experience during RLVR training? To answer this question, we propose \textbf{D}ual \textbf{G}uidance \textbf{O}ptimization~(\textbf{DGO}), a unified framework that leverages \emph{external} and \emph{internal experience} to improve training effectiveness. Specifically, DGO first constructs an experience bank from previously explored trajectories. The policy then performs exploration under the joint guidance of the experience bank and the model's internal knowledge. The resulting trajectories are further used to refine the experience bank and optimize model parameters, forming a closed loop of experience utilization and internalization. Experiments show that DGO consistently outperforms baseline methods, suggesting that better utilization and internalization of experience lead to more effective reasoning.

cs.LG↗

arXiv2026

Cross-Layer Semantic Flow Reconstruction for Attack Detection in Agentic Systems

Agentic systems increasingly orchestrate complex, tool-using workflows within agentic execution environments, where high-level goals and tool invocations at the application layer materialize as process, file, and network activities at the operating-system layer. This cross-layer execution creates security risks that conventional input guardrails cannot capture, because malicious intent may become observable only through downstream execution effects. In multi-agent deployments, inter-agent communication and delegation introduce additional propagation paths. To address this gap, we propose AScope, an execution-aware framework that correlates application-level agent semantics with kernel-level audit events and reconstructs them as cross-layer semantic flows. AScope connects fragmented operations into causal behavioral trajectories and uses a supervisor LLM to identify data flow violations, control flow deviations, and intent inconsistencies. We evaluate AScope on published AgentDojo traces with application-layer evidence and on ten multi-agent scenarios with cross-layer telemetry. The results demonstrate strong detection sensitivity across both evidence settings and achieve node- and path-level F1-scores of 85.3% and 66.7% on the cross-layer dataset.

cs.CR↗

arXiv2026

ForesightKV: Optimizing KV Cache Eviction for Reasoning Models by Learning Long-Term Contribution

Recently, large language models (LLMs) have shown remarkable reasoning abilities by producing long reasoning traces. However, as the sequence length grows, the key-value (KV) cache expands linearly, incurring significant memory and computation costs. Existing KV cache eviction methods mitigate this issue by discarding less important KV pairs, but often fail to capture complex KV dependencies, resulting in performance degradation. To better balance efficiency and performance, we introduce ForesightKV, a training-based KV cache eviction framework that learns to predict which KV pairs to evict during long-text generations. We first design the Golden Eviction algorithm, which identifies the optimal eviction KV pairs at each step using future attention scores. These traces and the scores at each step are then distilled via supervised training with a Pairwise Ranking Loss. Furthermore, we formulate cache eviction as a Markov Decision Process and apply the GRPO algorithm to mitigate the significant language modeling loss increase on low-entropy tokens. Experiments on AIME2024 and AIME2025 benchmarks of three reasoning models demonstrate that ForesightKV consistently outperforms prior methods under only half the cache budget, while benefiting synergistically from both supervised and reinforcement learning approaches. Code is available at https://github.com/RUCAIBox/ForesightKV.

cs.CL↗

arXiv2026

Adaptive Ability Decomposing for Unlocking Large Reasoning Model Effective Reinforcement Learning

Reinforcement learning with verifiable rewards (RLVR) has shown great potential to enhance the reasoning ability of large language models (LLMs). However, due to the limited amount of information provided during the RLVR process, the model can only engage in largely blind exploration, which often results in failure on challenging problems. To provide additional information for the RLVR process without relying on a teacher model, we propose A$^2$D, an Adaptive Ability Decomposing method for enhancing the effectiveness of RLVR. Specifically, we first train a decomposer via RLVR without distillation, enabling it to decompose complex questions into a set of simpler sub-questions. Next, we use this decomposer to annotate sub-questions for each question in the training dataset, and then train the reasoner under RLVR with sub-question guidance. To better understand A$^2$D, we first compare its performance with competitive baselines, showing its effectiveness. Next, we observe that our method functions as a plug-and-play module that can be applied to different RLVR algorithms. Furthermore, we conduct an analysis of the decomposer, revealing how the RLVR process affects its performance and behavior, and which type of guidance is better suited for enhancing the reasoner's exploration and exploitation abilities.

cs.CL↗

arXiv2026

PerfGuard: A Performance-Aware Agent for Visual Content Generation

The advancement of Large Language Model (LLM)-powered agents has enabled automated task processing through reasoning and tool invocation capabilities. However, existing frameworks often operate under the idealized assumption that tool executions are invariably successful, relying solely on textual descriptions that fail to distinguish precise performance boundaries and cannot adapt to iterative tool updates. This gap introduces uncertainty in planning and execution, particularly in domains like visual content generation (AIGC), where nuanced tool performance significantly impacts outcomes. To address this, we propose PerfGuard, a performance-aware agent framework for visual content generation that systematically models tool performance boundaries and integrates them into task planning and scheduling. Our framework introduces three core mechanisms: (1) Performance-Aware Selection Modeling (PASM), which replaces generic tool descriptions with a multi-dimensional scoring system based on fine-grained performance evaluations; (2) Adaptive Preference Update (APU), which dynamically optimizes tool selection by comparing theoretical rankings with actual execution rankings; and (3) Capability-Aligned Planning Optimization (CAPO), which guides the planner to generate subtasks aligned with performance-aware strategies. Experimental comparisons against state-of-the-art methods demonstrate PerfGuard's advantages in tool selection accuracy, execution reliability, and alignment with user intent, validating its robustness and practical utility for complex AIGC tasks. The project code is available at https://github.com/FelixChan9527/PerfGuard.

cs.AI↗

arXiv2026

LPIPS-AttnWav2Lip: Generic Audio-Driven lip synchronization for Talking Head Generation in the Wild

Researchers have shown a growing interest in Audio-driven Talking Head Generation. The primary challenge in talking head generation is achieving audio-visual coherence between the lips and the audio, known as lip synchronization. This paper proposes a generic method, LPIPS-AttnWav2Lip, for reconstructing face images of any speaker based on audio. We used the U-Net architecture based on residual CBAM to better encode and fuse audio and visual modal information. Additionally, the semantic alignment module extends the receptive field of the generator network to obtain the spatial and channel information of the visual features efficiently; and match statistical information of visual features with audio latent vector to achieve the adjustment and injection of the audio content information to the visual information. To achieve exact lip synchronization and to generate realistic high-quality images, our approach adopts LPIPS Loss, which simulates human judgment of image quality and reduces instability possibility during the training process. The proposed method achieves outstanding performance in terms of lip synchronization accuracy and visual quality as demonstrated by subjective and objective evaluation results. The code for the paper is available at the following link: https://github.com/FelixChan9527/LPIPS-AttnWav2Lip

cs.SD↗

arXiv2026

Formalization of Amicable Numbers Theory

This paper presents a formalization of the theory of amicable numbers in the Lean~4 proof assistant. Two positive integers $m$ and $n$ are called an amicable pair if the sum of proper divisors of $m$ equals $n$ and the sum of proper divisors of $n$ equals $m$. Our formalization introduces the proper divisor sum function $\propersum(n) = \sigma(n) - n$, defines the concepts of amicable pairs and amicable numbers, and computationally verifies historically famous amicable pairs. Furthermore, we formalize basic structural theorems, including symmetry, non-triviality, and connections to abundant/deficient numbers. A key contribution is the complete formal proof of the classical Th\={a}bit formula (9th century), using index-shifting and the \texttt{zify} tactic. Additionally, we provide complete formal proofs of both Th\={a}bit's rule and Euler's generalized rule (1747), two fundamental theorems for generating amicable pairs. A major achievement is the first complete formalization of the Borho-Hoffmann breeding method (1986), comprising 540 lines with 33 theorems and leveraging automated algebra tactics (\texttt{zify} and \texttt{ring}) to verify complex polynomial identities. We also formalize extensions including sociable numbers (aliquot cycles), betrothed numbers (quasi-amicable pairs), parity constraint theorems, and computational search bounds for coprime pairs ($>10^{65}$). We verify the smallest sociable cycle of length 5 (Poulet's cycle) and computationally verify specific instances. The formalization comprises 2076 lines of Lean code organized into Mathlib-candidate and paper-specific modules, with 139 theorems and all necessary infrastructure for divisor sum multiplicativity and coprimality reasoning.

cs.LO↗

arXiv2025

Pass@k Training for Adaptively Balancing Exploration and Exploitation of Large Reasoning Models

Reinforcement learning with verifiable rewards (RLVR), which typically adopts Pass@1 as the reward, has faced the issues in balancing exploration and exploitation, causing policies to prefer conservative actions, converging to a local optimum. Identifying an appropriate reward metric is therefore crucial. Regarding the prior work, although Pass@k has been used in evaluation, its connection to LLM exploration ability in RLVR remains largely overlooked. To investigate this, we first use Pass@k as the reward to train the policy model (i.e., $\textbf{Pass@k Training}$), and observe the improvement on its exploration ability. Next, we derive an analytical solution for the advantage of Pass@k Training, leading to an efficient and effective process. Building on this, our analysis reveals that exploration and exploitation are not inherently conflicting objectives, while they can mutually enhance each other. Moreover, Pass@k Training with analytical derivation essentially involves directly designing the advantage function. Inspired by this, we preliminarily explore the advantage design for RLVR, showing promising results and highlighting a potential future direction.

cs.LG↗

arXiv2025

From Trial-and-Error to Improvement: A Systematic Analysis of LLM Exploration Mechanisms in RLVR

Reinforcement learning with verifiable rewards (RLVR) has emerged as a powerful paradigm for enhancing the reasoning capabilities of large language models (LLMs). Unlike traditional RL approaches, RLVR leverages rule-based feedback to guide LLMs in generating and refining complex reasoning chains -- a process critically dependent on effective exploration strategies. While prior work has demonstrated RLVR's empirical success, the fundamental mechanisms governing LLMs' exploration behaviors remain underexplored. This technical report presents a systematic investigation of exploration capacities in RLVR, covering four main aspects: (1) exploration space shaping, where we develop quantitative metrics to characterize LLMs' capability boundaries; (2) entropy-performance exchange, analyzed across training stages, individual instances, and token-level patterns; and (3) RL performance optimization, examining methods to effectively translate exploration gains into measurable improvements. By unifying previously identified insights with new empirical evidence, this work aims to provide a foundational framework for advancing RLVR systems.

cs.CL↗

arXiv2025

HGMF: A Hierarchical Gaussian Mixture Framework for Scalable Tool Invocation within the Model Context Protocol

Invoking external tools enables Large Language Models (LLMs) to perform complex, real-world tasks, yet selecting the correct tool from large, hierarchically-structured libraries remains a significant challenge. The limited context windows of LLMs and noise from irrelevant options often lead to low selection accuracy and high computational costs. To address this, we propose the Hierarchical Gaussian Mixture Framework (HGMF), a probabilistic pruning method for scalable tool invocation. HGMF first maps the user query and all tool descriptions into a unified semantic space. The framework then operates in two stages: it clusters servers using a Gaussian Mixture Model (GMM) and filters them based on the query's likelihood. Subsequently, it applies the same GMM-based clustering and filtering to the tools associated with the selected servers. This hierarchical process produces a compact, high-relevance candidate set, simplifying the final selection task for the LLM. Experiments on a public dataset show that HGMF significantly improves tool selection accuracy while reducing inference latency, confirming the framework's scalability and effectiveness for large-scale tool libraries.

cs.AI↗

arXiv2025

Decomposing the Entropy-Performance Exchange: The Missing Keys to Unlocking Effective Reinforcement Learning

Recently, reinforcement learning with verifiable rewards (RLVR) has been widely used for enhancing the reasoning abilities of large language models (LLMs). A core challenge in RLVR involves managing the exchange between entropy and performance of policies. Despite the importance of this exchange, a fine-grained understanding of when and how this exchange operates most effectively remains limited. To bridge this gap, we conduct a systematic empirical analysis of the entropy-performance exchange mechanism of RLVR across different levels of granularity. Specifically, we first divide the training process into two distinct stages based on entropy dynamics, i.e., rising stage and plateau stage, and then systematically investigate how this mechanism varies across stage-level, instance-level, and token-level granularitiess. Our analysis reveals that, in the rising stage, entropy reduction in negative samples facilitates the learning of effective reasoning patterns, which in turn drives rapid performance gains. Moreover, in the plateau stage, learning efficiency strongly correlates with high-entropy tokens present in low-perplexity samples and those located at the end of sequences. Motivated by these findings, we propose two methods that dynamically adjust the reward signal using perplexity and positional information to focus RL updates on tokens that exhibit high learning potential, achieving improvements compared to the baseline methods on various LLMs.

cs.CL↗

arXiv2025

Effects of turbulence spreading and symmetry breaking on edge shear flow during sawtooth cycles in J-TEXT tokamak

Sawtooth oscillations can trigger off heat and turbulence pulses that propagate into the edge plasma, and thus enhancing the edge shear flow and inducing a transition from low confinement mode to high confinement mode. The influences of turbulence spreading and symmetry breaking on edge shear flow with sawtooth crashes are observed in the J-TEXT tokamak. The edge plasma turbulence and shear flow are measured using a fast reciprocating electrostatic probe array. After sawtooth crashes, the heat and turbulence pulses in the core propagate to the edge, with the turbulence pulse being faster than the heat pulse. After sawtooth crashes, the edge electron temperature increases and the edge turbulence is enhanced, with turbulence preceding temperature. The enhanced edge turbulence is mainly composed of two parts: the turbulence driven by local gradient and the turbulence spreading from core to edge. The development of the estimated turbulence spreading rate is prior to that of the turbulence driving rate. The increase in the turbulence intensity can cause the turbulent Reynold stress and its gradient to increase, thereby enhancing shear flows and radial electric fields. Turbulence spreading leads the edge Reynolds stresses to develop and the shear flow to be faster than edge electron temperature. The Reynolds stress arises from the symmetry breaking of the turbulence wave number spectrum. After sawtooth collapses, the joint probability density function of radial wave number and poloidal wave number of turbulence intensity exhibits strong asymmetry. These results show that the turbulence spreading and symmetry breaking can enhance turbulent Reynolds stress, thereby driving shear flows, after sawtooth has crashed.

physics.plasm-ph↗

arXiv2025

ICPC-Eval: Probing the Frontiers of LLM Reasoning with Competitive Programming Contests

With the significant progress of large reasoning models in complex coding and reasoning tasks, existing benchmarks, like LiveCodeBench and CodeElo, are insufficient to evaluate the coding capabilities of large language models (LLMs) in real competition environments. Moreover, current evaluation metrics such as Pass@K fail to capture the reflective abilities of reasoning models. To address these challenges, we propose \textbf{ICPC-Eval}, a top-level competitive coding benchmark designed to probing the frontiers of LLM reasoning. ICPC-Eval includes 118 carefully curated problems from 11 recent ICPC contests held in various regions of the world, offering three key contributions: 1) A challenging realistic ICPC competition scenario, featuring a problem type and difficulty distribution consistent with actual contests. 2) A robust test case generation method and a corresponding local evaluation toolkit, enabling efficient and accurate local evaluation. 3) An effective test-time scaling evaluation metric, Refine@K, which allows iterative repair of solutions based on execution feedback. The results underscore the significant challenge in evaluating complex reasoning abilities: top-tier reasoning models like DeepSeek-R1 often rely on multi-turn code feedback to fully unlock their in-context reasoning potential when compared to non-reasoning counterparts. Furthermore, despite recent advancements in code generation, these models still lag behind top-performing human teams. We release the benchmark at: https://github.com/RUCAIBox/Slow_Thinking_with_LLMs

cs.CL↗