Searcharxiv⌕ Search

arXiv subjects

Yunpeng Zhai

Publications and source records attributed to Yunpeng Zhai.

At least 19 recordsLinked to original sources
arXiv2026

RLCSD: Reinforcement Learning with Contrastive On-Policy Self-Distillation

On-policy self-distillation (OPSD) provides dense, token-level supervision for reasoning models by aligning a model's own distribution with that under privileged context, typically a verified solution. However, we show that the resulting distributional gap concentrates on style tokens rather than task-bearing ones, as the hinted model tends to produce shorter, more direct outputs. We term this pathology \emph{privilege-induced style drift}, which can destabilize training and shorten responses. To address this, we propose \textbf{RLCSD} (Reinforcement Learning with Contrastive on-policy Self-Distillation), which mitigates this drift by contrasting the teacher-student gap under a correct hint against that under a wrong hint, suppressing style shifts induced by hints regardless of correctness and yielding a signal more concentrated on task-bearing tokens. Experiments on Qwen3 (1.7B/4B/8B) and Olmo-3-7B-Think across mathematical and logical reasoning show that RLCSD consistently outperforms GRPO and prior OPSD methods, with additional results on agentic tasks supporting broader applicability. We further show that the contrastive principle is general: it plugs into existing OPSD methods to improve them, and its underlying insight extends to broader cross-model on-policy distillation.

cs.LG↗
arXiv2026

Omni-SafetyBench: A Benchmark for Safety Evaluation of Audio-Visual Large Language Models

Omni-modal Large Language Models (OLLMs) that integrate visual, auditory, and textual processing face severe safety risks. They exhibit fragile defenses against audio-visual joint harmful inputs and demonstrate inconsistent safety performance across different modalities, enabling simple modality-switching jailbreaks. However, existing safety benchmarks fail to comprehensively assess these risks due to the absence of audio-visual joint samples, limited modality coverage, and lack of parallel test cases for cross-modal consistency evaluation. To address these gaps, we introduce Omni-SafetyBench, the first comprehensive parallel benchmark for OLLM safety evaluation, featuring 23,328 test instances across 24 modality variations derived from 972 seed samples. Recognizing that complex inputs pose comprehension challenges and that cross-modal consistency is critical for OLLM safety, we propose tailored metrics: a Safety-score based on Conditional Attack Success Rate (C-ASR) and Conditional Refusal Rate (C-RR), and a Cross-Modal Safety Consistency score (CMSC-score). Evaluating 11 state-of-the-art OLLMs reveals severe vulnerabilities: only 3 models exceed 0.6 in both metrics, with safety degrading sharply for audio-visual inputs. Furthermore, evaluation of existing safety alignment methods on Omni-SafetyBench identifies fundamental challenges in OLLM safety alignment, highlighting urgent needs for enhanced research in this domain.

cs.CL↗
arXiv2026

From Feedback Loops to Policy Updates: Reinforcement Fine-Tuning for LLM-Based Alpha Factor Discovery

Modern quantitative trading increasingly relies on systematic models to extract predictive signals from large-scale financial data, where alpha factor discovery plays a central role in transforming market observations into tradable signals. Recent LLM-based methods have shown promise in automating factor generation, but most of them still rely on prompt-level generation--evaluation--feedback loops for iterative optimization. As the loop becomes longer, repeatedly appended historical candidates and feedback can cause context explosion, increase inference cost, dilute useful information, and introduce feedback drift. Moreover, these methods often depend on very large LLMs whose stable generation preferences may lead to structurally similar expressions, redundant candidates, and search stagnation. To address these limitations, we propose \textsc{QuantEvolver}, a self-evolving alpha factor discovery framework based on reinforcement fine-tuning. Instead of accumulating feedback in the prompt, \textsc{QuantEvolver} converts executable quantitative evaluation into policy updates, enabling a Miner LLM to internalize historical optimization experience through parameter learning. Specifically, \textsc{QuantEvolver} constructs high-quality seed factors, builds diverse seed--time-window training tasks, generates executable Factor DSL expressions, evaluates them through Regime Backtest, and optimizes the Miner LLM with Diversity-Complementarity Reward. During training, high-quality factors are continuously accumulated in a Mined Factor Database, which serves as the final discovered factor library. Extensive experiments on three realistic market benchmarks demonstrate the effectiveness of \textsc{QuantEvolver}, which consistently improves the primary evaluation metric of each task over existing LLM-based alpha factor discovery baselines, produces higher-quality and more complementary factor pools.

cs.CE↗
arXiv2026

d-TreeRPO: Towards More Reliable Policy Optimization for Diffusion Language Models

Reinforcement learning (RL) is pivotal for enhancing the reasoning capabilities of diffusion large language models (dLLMs). However, existing dLLM policy optimization methods suffer from two critical reliability bottlenecks: (1) reward sparsity, arising from coarse or unverifiable signals that impede accurate advantage calculation; and (2) their probability estimates do not account for the gap to the unbiased expectation over all decoding orders, which are intractable to compute. To mitigate these issues, we propose d-TreeRPO, a reliable RL framework for dLLMs that leverages tree-structured rollouts and bottom-up advantage computation based on verifiable outcome rewards to provide fine-grained and verifiable step-wise reward signals. Furthermore, we provide a theoretical proof demonstrating that increasing prediction confidence effectively minimizes the gap between unbiased expected prediction probabilities and its single-step forward pass estimate. Guided by this analysis, we introduce a time-scheduled self-distillation loss during training that enhances prediction confidence in later training stages, thereby enabling more accurate probability estimation and better performance. Experiments demonstrate that d-TreeRPO outperforms existing baselines and achieves significant improvements across multiple reasoning benchmarks. Specifically, it achieves +86.2% on Sudoku, +51.6% on Countdown, +4.5% on GSM8K, and +5.3% on Math500 compared to the base model.

cs.CL↗
arXiv2026

Towards Robust LLM Post-Training: Automatic Failure Management for Reinforcement Fine-Tuning

Reinforcement fine-tuning (RFT) has become a core paradigm for post-training large language models, yet its training process remains highly fragile. Existing efforts mainly improve reliability at the system level or address specific issues in individual subproblems by modifying RFT algorithms. Despite their effectiveness, they largely overlook the problem of failure management at the training-process level. When training goes wrong, practitioners still rely heavily on expert-driven manual inspection and correction, and automatic failure management for RFT remains largely unexplored. In this paper, we take a first step toward systematic failure management for reinforcement fine-tuning. To understand the empirical structure of RFT failures, we first construct RFT-FaultBench, the first benchmark for fine-grained failures in reinforcement fine-tuning, covering 5 fault families, 16 fault types, 779 training runs, 22,549 train-step records, and 1,457,288 trajectory-level records. Based on this benchmark, we conduct a comprehensive empirical study showing that RFT failures are both observable from training dynamics and distinguishable through their empirical fault fingerprints. Building on these findings, we propose RFT-FM, an automatic failure management framework for reinforcement fine-tuning that unifies anomaly detection, failure diagnosis, and auto remediation in a closed loop. Experimental results show that RFT-FaultBench is neither trivial nor saturated: it exhibits clear anomaly structure while still posing substantial challenges, especially under subtle fault settings. Moreover, RFT-FM shows strong capability in detecting, diagnosing, and mitigating RFT failures.

cs.SE↗
arXiv2026

E2E-REME: Towards End-to-End Microservices Auto-Remediation via Experience-Simulation Reinforcement Fine-Tuning

Contemporary microservice systems continue to grow in scale and complexity, leading to increasingly frequent and costly failures. While recent LLM-based auto-remediation approaches have emerged, they primarily translate textual instructions into executable Ansible playbooks and rely on expert-crafted prompts, lacking runtime knowledge guidance and depending on large-scale general-purpose LLMs, which limits their accuracy and efficiency. We introduce \textit{End-to-End Microservice Remediation} (E2E-MR), a new task that requires directly generating executable playbooks from diagnosis reports to autonomously restore faulty systems. To enable rigorous evaluation, we build \textit{MicroRemed}, a benchmark that automates microservice deployment, failure injection, playbook execution, and post-repair verification. We further propose \textit{E2E-REME}, an end-to-end auto-remediation model trained via experience-simulation reinforcement fine-tuning. Experiments on public and industrial microservice platforms, compared with nine representative LLMs, show that E2E-REME achieves superior accuracy and efficiency.

cs.SE↗
arXiv2026

A Survey on Parallel Text Generation: From Parallel Decoding to Diffusion Language Models

As text generation has become a core capability of modern Large Language Models (LLMs), it underpins a wide range of downstream applications. However, most existing LLMs rely on autoregressive (AR) generation, producing one token at a time based on previously generated context-resulting in limited generation speed due to the inherently sequential nature of the process. To address this challenge, an increasing number of researchers have begun exploring parallel text generation-a broad class of techniques aimed at breaking the token-by-token generation bottleneck and improving inference efficiency. Despite growing interest, there remains a lack of comprehensive analysis on what specific techniques constitute parallel text generation and how they improve inference performance. To bridge this gap, we present a systematic survey of parallel text generation methods. We categorize existing approaches into AR-based and Non-AR-based paradigms, and provide a detailed examination of the core techniques within each category. Following this taxonomy, we assess their theoretical trade-offs in terms of speed, quality, and efficiency, and examine their potential for combination and comparison with alternative acceleration strategies. Finally, based on our findings, we highlight recent advancements, identify open challenges, and outline promising directions for future research in parallel text generation. We have also created a GitHub repository for indexing relevant papers and open resources available at https://github.com/zhanglingzhe0820/Awesome-Parallel-Text-Generation.

cs.CL↗
arXiv2026

Auto-Rubric: Learning From Implicit Weights to Explicit Rubrics for Reward Modeling

Conventional reward modeling relies on gradient descent over neural weights, creating opaque, data-hungry "black boxes." We propose a paradigm shift from implicit to explicit reward parameterization, recasting optimization from continuous weight spaces to the discrete space of natural language rubrics. We introduce a training-free framework based on iterative rubric learning: it locally induces discriminative criteria via verification-driven refinement, and globally compresses the candidate criteria pool into a compact core set by maximizing an information-theoretic coding rate objective. We organize the compressed core set into a hierarchical rubric structure -- high-level evaluation dimensions supported by concrete verification checks -- serving as an interpretable, portable reward function. Empirically, our approach challenges prevailing data scaling assumptions: using only 70 preference pairs, our rubric-guided judges outperform fully trained reward models on diverse benchmarks. For instance, Qwen3-8B equipped with our learned rubrics achieves 80.91% on RewardBench2, surpassing the specialized Skywork-Reward-V2-Qwen3-8B (78.20%). These results demonstrate that alignment signals are highly compressible and can be effectively captured through explicit symbolic search.

cs.LG↗
arXiv2026

ThinkFL: Self-Refining Failure Localization for Microservice Systems via Reinforcement Fine-Tuning

As modern microservice systems grow increasingly popular and complex-often consisting of hundreds or even thousands of fine-grained, interdependent components-they are becoming more susceptible to frequent and subtle failures. Ensuring system reliability therefore hinges on accurate and efficient failure localization. Traditional failure localization approaches based on small models lack the flexibility to adapt to diverse failure scenarios, while recent LLM-based methods suffer from two major limitations: they often rely on rigid invocation workflows that constrain the model's ability to dynamically explore optimal localization paths, and they require resource-intensive inference, making them cost-prohibitive for real-world deployment. To address these challenges, we explore the use of reinforcement fine-tuning to equip lightweight LLMs with reasoning and self-refinement capabilities, significantly improving the cost-effectiveness and adaptability of LLM-based failure localization. We begin with an empirical study to identify three key capabilities essential for accurate localization. Building on these insights, we propose a progressive multi-stage GRPO fine-tuning framework, which integrates a multi-factor failure localization grader and a recursion-of-thought actor module. The resulting model, ThinkFL, not only outperforms existing state-of-the-art LLMs and baseline methods in localization accuracy but also reduces end-to-end localization latency from minutes to seconds, demonstrating strong potential for real-world applications.

cs.SE↗
arXiv2026

Hypothesize-Then-Verify: Speculative Root Cause Analysis for Microservices with Pathwise Parallelism

Microservice systems have become the backbone of cloud-native enterprise applications due to their resource elasticity, loosely coupled architecture, and lightweight deployment. Yet, the intrinsic complexity and dynamic runtime interactions of such systems inevitably give rise to anomalies. Ensuring system reliability therefore hinges on effective root cause analysis (RCA), which entails not only localizing the source of anomalies but also characterizing the underlying failures in a timely and interpretable manner. Recent advances in intelligent RCA techniques, particularly those powered by large language models (LLMs), have demonstrated promising capabilities, as LLMs reduce reliance on handcrafted features while offering cross-platform adaptability, task generalization, and flexibility. However, existing LLM-based methods still suffer from two critical limitations: (a) limited exploration diversity, which undermines accuracy, and (b) heavy dependence on large-scale LLMs, which results in slow inference. To overcome these challenges, we propose SpecRCA, a speculative root cause analysis framework for microservices that adopts a \textit{hypothesize-then-verify} paradigm. SpecRCA first leverages a hypothesis drafting module to rapidly generate candidate root causes, and then employs a parallel root cause verifier to efficiently validate them. Preliminary experiments on the AIOps 2022 dataset demonstrate that SpecRCA achieves superior accuracy and efficiency compared to existing approaches, highlighting its potential as a practical solution for scalable and interpretable RCA in complex microservice environments.

cs.SE↗
arXiv2026

Agentic Memory Enhanced Recursive Reasoning for Root Cause Localization in Microservices

As contemporary microservice systems become increasingly popular and complex-often comprising hundreds or even thousands of fine-grained, interdependent subsystems-they are experiencing more frequent failures. Ensuring system reliability thus demands accurate root cause localization. While many traditional graph-based and deep learning approaches have been explored for this task, they often rely heavily on pre-defined schemas that struggle to adapt to evolving operational contexts. Consequently, a number of LLM-based methods have recently been proposed. However, these methods still face two major limitations: shallow, symptom-centric reasoning that undermines accuracy, and a lack of cross-alert reuse that leads to redundant reasoning and high latency. In this paper, we conduct a comprehensive study of how Site Reliability Engineers (SREs) localize the root causes of failures, drawing insights from professionals across multiple organizations. Our investigation reveals that expert root cause analysis exhibits three key characteristics: recursiveness, multi-dimensional expansion, and cross-modal reasoning. Motivated by these findings, we introduce AMER-RCL, an agentic memory enhanced recursive reasoning framework for root cause localization in microservices. AMER-RCL employs the Recursive Reasoning RCL engine, a multi-agent framework that performs recursive reasoning on each alert to progressively refine candidate causes, while Agentic Memory incrementally accumulates and reuses reasoning from prior alerts within a time window to reduce redundant exploration and lower inference latency. Experimental results demonstrate that AMER-RCL consistently outperforms state-of-the-art methods in both localization accuracy and inference efficiency.

cs.SE↗
arXiv2025

CuES: A Curiosity-driven and Environment-grounded Synthesis Framework for Agentic RL

Large language model based agents are increasingly deployed in complex, tool augmented environments. While reinforcement learning provides a principled mechanism for such agents to improve through interaction, its effectiveness critically depends on the availability of structured training tasks. In many realistic settings, however, no such tasks exist a challenge we term task scarcity, which has become a key bottleneck for scaling agentic RL. Existing approaches typically assume predefined task collections, an assumption that fails in novel environments where tool semantics and affordances are initially unknown. To address this limitation, we formalize the problem of Task Generation for Agentic RL, where an agent must learn within a given environment that lacks predefined tasks. We propose CuES, a Curiosity driven and Environment grounded Synthesis framework that autonomously generates diverse, executable, and meaningful tasks directly from the environment structure and affordances, without relying on handcrafted seeds or external corpora. CuES drives exploration through intrinsic curiosity, abstracts interaction patterns into reusable task schemas, and refines them through lightweight top down guidance and memory based quality control. Across three representative environments, AppWorld, BFCL, and WebShop, CuES produces task distributions that match or surpass manually curated datasets in both diversity and executability, yielding substantial downstream policy improvements. These results demonstrate that curiosity driven, environment grounded task generation provides a scalable foundation for agents that not only learn how to act, but also learn what to learn. The code is available at https://github.com/modelscope/AgentEvolver/tree/main/research/CuES.

cs.AI↗
arXiv2025

AgentEvolver: Towards Efficient Self-Evolving Agent System

Autonomous agents powered by large language models (LLMs) have the potential to significantly enhance human productivity by reasoning, using tools, and executing complex tasks in diverse environments. However, current approaches to developing such agents remain costly and inefficient, as they typically require manually constructed task datasets and reinforcement learning (RL) pipelines with extensive random exploration. These limitations lead to prohibitively high data-construction costs, low exploration efficiency, and poor sample utilization. To address these challenges, we present AgentEvolver, a self-evolving agent system that leverages the semantic understanding and reasoning capabilities of LLMs to drive autonomous agent learning. AgentEvolver introduces three synergistic mechanisms: (i) self-questioning, which enables curiosity-driven task generation in novel environments, reducing dependence on handcrafted datasets; (ii) self-navigating, which improves exploration efficiency through experience reuse and hybrid policy guidance; and (iii) self-attributing, which enhances sample efficiency by assigning differentiated rewards to trajectory states and actions based on their contribution. By integrating these mechanisms into a unified framework, AgentEvolver enables scalable, cost-effective, and continual improvement of agent capabilities. Preliminary experiments indicate that AgentEvolver achieves more efficient exploration, better sample utilization, and faster adaptation compared to traditional RL-based baselines.

cs.LG↗
arXiv2025

MicroRemed: Benchmarking LLMs in Microservices Remediation

Large Language Models (LLMs) integrated with agent-based reasoning frameworks have recently shown strong potential for autonomous decision-making and system-level operations. One promising yet underexplored direction is microservice remediation, where the goal is to automatically recover faulty microservice systems. Existing approaches, however, still rely on human-crafted prompts from Site Reliability Engineers (SREs), with LLMs merely converting textual instructions into executable code. To advance research in this area, we introduce MicroRemed, the first benchmark for evaluating LLMs in end-to-end microservice remediation, where models must directly generate executable Ansible playbooks from diagnosis reports to restore system functionality. We further propose ThinkRemed, a multi-agent framework that emulates the reflective and perceptive reasoning of SREs. Experimental results show that MicroRemed presents substantial challenges to current LLMs, while ThinkRemed improves end-to-end remediation performance through iterative reasoning and system reflection. The benchmark is available at https://github.com/LLM4AIOps/MicroRemed.

cs.CL↗
arXiv2025

A Fully-integrated Diamond Nitrogen-Vacancy Magnetometer with Nanotesla Sensitivity

Ensemble diamond nitrogen-vacancy (DNV) centers have emerged as a promising platform for precise earth-field vector magnetic sensing, particularly in applications that require high mobility. Nevertheless, integrating all control utilities into a compact form has proven challenging, thus far limiting the sensitivity of mobile DNV magnetometers to the uT-level. This study introduces a fully integrated DNV magnetometer that encompasses all the essential components typically found in traditional platforms, while maintaining compact dimensions of approximately 13 cm * 26 cm. In contrast to previous efforts, we successfully address these challenges by integrating a high-power laser, a lock-in amplifier, and a digitally-modulated microwave source. These home-made components show comparable performance with commercial devices under our circumstance, resulting in an optimal sensitivity of 2.14 nT/sqrt{Hz}. The limitations in this system as well as possible future improvements are discussed. This work paves the way for the use of DNV magnetometry in cost-effective, mobile unmanned aerial vehicles, facilitating a wide range of practical applications.

quant-ph↗
arXiv2025

Provoking Multi-modal Few-Shot LVLM via Exploration-Exploitation In-Context Learning

In-context learning (ICL), a predominant trend in instruction learning, aims at enhancing the performance of large language models by providing clear task guidance and examples, improving their capability in task understanding and execution. This paper investigates ICL on Large Vision-Language Models (LVLMs) and explores the policies of multi-modal demonstration selection. Existing research efforts in ICL face significant challenges: First, they rely on pre-defined demonstrations or heuristic selecting strategies based on human intuition, which are usually inadequate for covering diverse task requirements, leading to sub-optimal solutions; Second, individually selecting each demonstration fails in modeling the interactions between them, resulting in information redundancy. Unlike these prevailing efforts, we propose a new exploration-exploitation reinforcement learning framework, which explores policies to fuse multi-modal information and adaptively select adequate demonstrations as an integrated whole. The framework allows LVLMs to optimize themselves by continually refining their demonstrations through self-exploration, enabling the ability to autonomously identify and generate the most effective selection policies for in-context learning. Experimental results verify the superior performance of our approach on four Visual Question-Answering (VQA) datasets, demonstrating its effectiveness in enhancing the generalization capability of few-shot LVLMs.

cs.CV↗
arXiv2025

Contrastive Learning for Continuous Touch-Based Authentication

Smart mobile devices have become indispensable in modern daily life, where sensitive information is frequently processed, stored, and transmitted-posing critical demands for robust security controls. Given that touchscreens are the primary medium for human-device interaction, continuous user authentication based on touch behavior presents a natural and seamless security solution. While existing methods predominantly adopt binary classification under single-modal learning settings, we propose a unified contrastive learning framework for continuous authentication in a non-disruptive manner. Specifically, the proposed method leverages a Temporal Masked Autoencoder to extract temporal patterns from raw multi-sensor data streams, capturing continuous motion and gesture dynamics. The pre-trained TMAE is subsequently integrated into a Siamese Temporal-Attentive Convolutional Network within a contrastive learning paradigm to model both sequential and cross-modal patterns. To further enhance performance, we incorporate multi-head attention and channel attention mechanisms to capture long-range dependencies and optimize inter-channel feature integration. Extensive experiments on public benchmarks and a self-collected dataset demonstrate that our approach outperforms state-of-the-art methods, offering a reliable and effective solution for user authentication on mobile devices.

cs.CR↗
arXiv2025

AgentFM: Role-Aware Failure Management for Distributed Databases with LLM-Driven Multi-Agents

Distributed databases are critical infrastructures for today's large-scale software systems, making effective failure management essential to ensure software availability. However, existing approaches often overlook the role distinctions within distributed databases and rely on small-scale models with limited generalization capabilities. In this paper, we conduct a preliminary empirical study to emphasize the unique significance of different roles. Building on this insight, we propose AgentFM, a role-aware failure management framework for distributed databases powered by LLM-driven multi-agents. AgentFM addresses failure management by considering system roles, data roles, and task roles, with a meta-agent orchestrating these components. Preliminary evaluations using Apache IoTDB demonstrate the effectiveness of AgentFM and open new directions for further research.

cs.SE↗