Yunpeng Zhai

Yunpeng Zhai, Peixi Peng, Mengxi Jia et al. · pku

Unsupervised person re-identification has achieved great success through the self-improvement of individual neural networks. However, limited by the lack of diversity of discriminant information, a single network has difficulty learning sufficient discrimination ability by itself under unsupervised conditions. To address this limit, we develop a population-based evolutionary gaming (PEG) framework in which a population of diverse neural networks is trained concurrently through selection, reproduction, mutation, and population mutual learning iteratively. Specifically, the selection of networks to preserve is modeled as a cooperative game and solved by the best-response dynamics, then the reproduction and mutation are implemented by cloning and fluctuating hyper-parameters of networks to learn more diversity, and population mutual learning improves the discrimination of networks by knowledge distillation from each other within the population. In addition, we propose a cross-reference scatter (CRS) to approximately evaluate re-ID models without labeled samples and adopt it as the criterion of network selection in PEG. CRS measures a model's performance by indirectly estimating the accuracy of its predicted pseudo-labels according to the cohesion and separation of the feature space. Extensive experiments demonstrate that (1) CRS approximately measures the performance of models without labeled samples; (2) and PEG produces new state-of-the-art accuracy for person re-identification, indicating the great potential of population-based network cooperative training for unsupervised learning.

Lingzhe Zhang, Tong Jia, Yunpeng Zhai et al.

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.

Lingzhe Zhang, Yunpeng Zhai, Tong Jia et al.

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.

Shinji Mai, Yunpeng Zhai, Ziqian Chen et al.

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/research/CuES.

Lingzhe Zhang, Yunpeng Zhai, Tong Jia et al.

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.

Leyi Pan, Shuchang Tao, Yunpeng Zhai et al.

Reliable reinforcement learning (RL) for diffusion large language models (dLLMs) requires both accurate advantage estimation and precise estimation of prediction probabilities. Existing RL methods for dLLMs fall short in both aspects: they rely on coarse or unverifiable reward signals, and they estimate prediction probabilities without accounting for the bias relative to the true, unbiased expected prediction probability that properly integrates over all possible decoding orders. To mitigate these issues, we propose \emph{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. When estimating the conditional transition probability from a parent node to a child node, we theoretically analyze the estimation error between the unbiased expected prediction probability and the estimate obtained via a single forward pass, and find that higher prediction confidence leads to lower estimation error. 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 improved convergence. Experiments show that \emph{d}-TreeRPO outperforms existing baselines and achieves significant gains on multiple reasoning benchmarks, including +86.2 on Sudoku, +51.6 on Countdown, +4.5 on GSM8K, and +5.3 on Math500. Ablation studies and computational cost analyses further demonstrate the effectiveness and practicality of our design choices.

Yunpeng Zhai, Shuchang Tao, Cheng Chen et al.

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.

Lingzhe Zhang, Liancheng Fang, Chiming Duan et al. · tsinghua

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.

Lingzhe Zhang, Tong Jia, Yunpeng Zhai et al.

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.

Leyi Pan, Zheyu Fu, Yunpeng Zhai et al. · tsinghua

The rise of Omni-modal Large Language Models (OLLMs), which integrate visual and auditory processing with text, necessitates robust safety evaluations to mitigate harmful outputs. However, no dedicated benchmarks currently exist for OLLMs, and existing benchmarks fail to assess safety under joint audio-visual inputs or cross-modal consistency. To fill this gap, we introduce Omni-SafetyBench, the first comprehensive parallel benchmark for OLLM safety evaluation, featuring 24 modality variations with 972 samples each, including audio-visual harm cases. Considering OLLMs' comprehension challenges with complex omni-modal inputs and the need for cross-modal consistency evaluation, we propose tailored metrics: a Safety-score based on Conditional Attack Success Rate (C-ASR) and Refusal Rate (C-RR) to account for comprehension failures, and a Cross-Modal Safety Consistency score (CMSC-score) to measure consistency across modalities. Evaluating 6 open-source and 4 closed-source OLLMs reveals critical vulnerabilities: (1) only 3 models achieving over 0.6 in both average Safety-score and CMSC-score; (2) safety defenses weaken with complex inputs, especially audio-visual joints; (3) severe weaknesses persist, with some models scoring as low as 0.14 on specific modalities. Using Omni-SafetyBench, we evaluated existing safety alignment algorithms and identified key challenges in OLLM safety alignment: (1) Inference-time methods are inherently less effective as they cannot alter the model's underlying understanding of safety; (2) Post-training methods struggle with out-of-distribution issues due to the vast modality combinations in OLLMs; and, safety tasks involving audio-visual inputs are more complex, making even in-distribution training data less effective. Our proposed benchmark, metrics and the findings highlight urgent needs for enhanced OLLM safety.

Lipeng Xie, Sen Huang, Zhuo Zhang et al.

Reward models are essential for aligning Large Language Models (LLMs) with human values, yet their development is hampered by costly preference datasets and poor interpretability. While recent rubric-based approaches offer transparency, they often lack systematic quality control and optimization, creating a trade-off between scalability and reliability. We address these limitations with a novel, training-free framework built on a key assumption: \textit{evaluation rubrics underlying human preferences exhibit significant generalization ability across diverse queries}, a property that enables remarkable data efficiency. Our two-stage approach first infers high-quality, query-specific rubrics using a validation-guided \textbf{Propose-Evaluate-Revise} pipeline. Second, it generalizes these granular rubrics into a compact, non-redundant core set by maximizing an \textbf{information-theoretic coding rate}. The final output is an interpretable, hierarchical "Theme-Tips" rubric set. Extensive experiments demonstrate the framework's exceptional data efficiency and performance. Critically, using just 70 preference pairs (1.5\% of the source data), our method also empowers smaller models like Qwen3-8B to outperform specialized, fully-trained counterparts. This work pioneers a scalable, interpretable, and data-efficient path for reward modeling.

Cheng Chen, Yunpeng Zhai, Yifan Zhao et al. · pku

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.

Yunpeng Zhai, Qixiang Ye, Shijian Lu et al.

Often the best performing deep neural models are ensembles of multiple base-level networks, nevertheless, ensemble learning with respect to domain adaptive person re-ID remains unexplored. In this paper, we propose a multiple expert brainstorming network (MEB-Net) for domain adaptive person re-ID, opening up a promising direction about model ensemble problem under unsupervised conditions. MEB-Net adopts a mutual learning strategy, where multiple networks with different architectures are pre-trained within a source domain as expert models equipped with specific features and knowledge, while the adaptation is then accomplished through brainstorming (mutual learning) among expert models. MEB-Net accommodates the heterogeneity of experts learned with different architectures and enhances discrimination capability of the adapted re-ID model, by introducing a regularization scheme about authority of experts. Extensive experiments on large-scale datasets (Market-1501 and DukeMTMC-reID) demonstrate the superior performance of MEB-Net over the state-of-the-arts.

Mengxi Jia, Yunpeng Zhai, Shijian Lu et al.

RGB-Infrared (IR) cross-modality person re-identification (re-ID), which aims to search an IR image in RGB gallery or vice versa, is a challenging task due to the large discrepancy between IR and RGB modalities. Existing methods address this challenge typically by aligning feature distributions or image styles across modalities, whereas the very useful similarities among gallery samples of the same modality (i.e. intra-modality sample similarities) is largely neglected. This paper presents a novel similarity inference metric (SIM) that exploits the intra-modality sample similarities to circumvent the cross-modality discrepancy targeting optimal cross-modality image matching. SIM works by successive similarity graph reasoning and mutual nearest-neighbor reasoning that mine cross-modality sample similarities by leveraging intra-modality sample similarities from two different perspectives. Extensive experiments over two cross-modality re-ID datasets (SYSU-MM01 and RegDB) show that SIM achieves significant accuracy improvement but with little extra training as compared with the state-of-the-art.

Yunpeng Zhai, Shijian Lu, Qixiang Ye et al.

Domain adaptive person re-identification (re-ID) is a challenging task, especially when person identities in target domains are unknown. Existing methods attempt to address this challenge by transferring image styles or aligning feature distributions across domains, whereas the rich unlabeled samples in target domains are not sufficiently exploited. This paper presents a novel augmented discriminative clustering (AD-Cluster) technique that estimates and augments person clusters in target domains and enforces the discrimination ability of re-ID models with the augmented clusters. AD-Cluster is trained by iterative density-based clustering, adaptive sample augmentation, and discriminative feature learning. It learns an image generator and a feature encoder which aim to maximize the intra-cluster diversity in the sample space and minimize the intra-cluster distance in the feature space in an adversarial min-max manner. Finally, AD-Cluster increases the diversity of sample clusters and improves the discrimination capability of re-ID models greatly. Extensive experiments over Market-1501 and DukeMTMC-reID show that AD-Cluster outperforms the state-of-the-art with large margins.