11 papers
Before the Rollout Ends: Early Terminal Reward Prediction for Long-horizon Coding Agents
Jihan Yao, Sihan Zeng, Shangbin Feng +3
Long-horizon coding agents receive verifiable rewards only after completing expensive sequences of tool calls. This increases inference cost, amplifies early wrong hypotheses, and…
Environment Evolution for Terminal Agents
Zhiyuan Fan, Tinghao Yu, Yuanjun Cai +9
Scaling interactive and verifiable environments is critical for training terminal agents. As frontier models become more capable, environments synthesized from scratch become less…
Deep Research Pretraining via Predictive Navigation
Jiang Zhou, Zhiyuan Fan, Xing Wu +3
Deep research agents are often trained on expensive, environment-grounded tool-use trajectories that require repeated retrieval, document inspection, and report evaluation. We intr…
Evolving-RL: End-to-End Optimization of Experience-Driven Self-Evolving Capability within Agents
Zhiyuan Fan, Wenwei Jin, Feng Zhang +4
Experience-driven self-evolving agents aim to overcome the static nature of large language models by distilling reusable experience from past interactions, thus enabling adaptation…
Toward Scalable Terminal Task Synthesis via Skill Graphs
Zhiyuan Fan, Tinghao Yu, Yuanjun Cai +8
Terminal agents have demonstrated strong potential for autonomous command-line execution, yet their training remains constrained by the scarcity of high-quality and diverse executi…
Evaluating the Formal Reasoning Capabilities of Large Language Models through Chomsky Hierarchy
Yihong Dong, Jianha Xiao, Xue Jiang +7
The formal reasoning capabilities of LLMs are crucial for advancing automated software engineering. However, existing benchmarks for LLMs lack systematic evaluation based on comput…
PretrainZero: Reinforcement Active Pretraining
Xingrun Xing, Zhiyuan Fan, Jie Lou +3
Mimicking human behavior to actively learning from general experience and achieve artificial general intelligence has always been a human dream. Recent reinforcement learning (RL)…
Saber: An Efficient Sampling with Adaptive Acceleration and Backtracking Enhanced Remasking for Diffusion Language Model
Yihong Dong, Zhaoyu Ma, Xue Jiang +10
Diffusion language models (DLMs) are emerging as a compelling alternative to the dominant autoregressive paradigm, offering inherent advantages in parallel generation and bidirecti…
SWE-Effi: Re-Evaluating Software AI Agent System Effectiveness Under Resource Constraints
Zhiyu Fan, Kirill Vasilevski, Dayi Lin +6
The advancement of large language models (LLMs) and code agents has demonstrated significant potential to assist software engineering (SWE) tasks, such as autonomous issue resoluti…
SedarEval: Automated Evaluation using Self-Adaptive Rubrics
Zhiyuan Fan, Weinong Wang, Xing Wu +1
The evaluation paradigm of LLM-as-judge gains popularity due to its significant reduction in human labor and time costs. This approach utilizes one or more large language models (L…
Efficient Data Learning for Open Information Extraction with Pre-trained Language Models
Zhiyuan Fan, Shizhu He
Open Information Extraction (OpenIE) is a fundamental yet challenging task in Natural Language Processing, which involves extracting all triples (subject, predicate, object) from a…