13 citations · 32 across the 68 of their papers we have counts for
80 papers
Harness Self-Evolution for Long-Horizon Terminal Agents
Ruhan Wang, Yucheng Shi, Zongxia Li +6
Long-horizon terminal agents may solve a task in one run yet fail across repeated runs. Stochastic errors can propagate through actions and context, creating a gap between capabili…
Terminal-RSI: Recursive Self-Improving Terminal Agents via Recursive Self-Rewrite and Policy-Refine
Junyao Yang, Zongxia Li, Yucheng Shi +6
Converting verifier outcomes into useful policy updates and partial progress into complete task resolution is crucial for terminal agents. We present Terminal-RSI, which alternates…
Long-Horizon-Terminal-Bench 2.0: Automating Diverse Task Construction and Measuring Sustained Agent Progress
Zongxia Li, Zhongzhi Li, Yucheng Shi +14
Long-horizon terminal benchmarks measure whether an agent can sustain progress through extended workflows with process rewards, not only whether its final state passes a binary che…
Scaling Trajectories for Complex Tasks through Recursive Self-Rewrite
Zongxia Li, Yucheng Shi, Zhongzhi Li +7
Successful trajectories on difficult tasks provide valuable supervision for model improvement, but specialized harnesses introduce interventions that may be unavailable during depl…
Self-Supervised Scaling of Terminal Environments for Scientific Domains
Zhongzhi Li, Yucheng Shi, Zongxia Li +8
Terminal agents are increasingly deployed beyond software engineering in science and other specialized domains. Constructing training environments requires executable reference beh…
T1: Terminal Agent Reinforcement Learning for Long-Horizon Tasks
Junyao Yang, Yucheng Shi, Zhongzhi Li +4
Agent usage is shifting toward long-horizon tasks such as coding and scientific discovery, among which terminal tasks are especially important. We introduce T1, a Mixture-of-Expert…
Recursive Synthesis for Long-Horizon Terminal Tasks
Zhongzhi Li, Yucheng Shi, Zongxia Li +8
High-quality long-horizon training data for terminal agents is expensive to produce, often costing hundreds to thousands of dollars per task, because each task must keep the instru…
Stale but Stable: Staleness-Adaptive Trust Regions for Stabilizing Asynchronous Reinforcement Learning
Junyao Yang, Yucheng Shi, Zongxia Li +6
Asynchronous reinforcement learning improves throughput by decoupling rollout generation from optimization, but the resulting staleness is an inevitable byproduct, compounded joint…
FlashMemory-DeepSeek-V4: Lightning Index Ultra-Long Context via Lookahead Sparse Attention
Yan Wang, Qifan Zhang, Jiachen Yu +12
Conventional LLMs keep the full KV cache loaded during decoding, causing a severe GPU memory bottleneck for ultra-long context serving. In this report, we propose \textbf{Lookahead…
Harness Handbook: Making Evolving Agent Harnesses Readable,Navigable, and Editable
Ruhan Wang, Yucheng Shi, Zongxia Li +7
The capability of a modern AI agent depends not only on its foundation model but also on its harness, which constructs prompts, manages state, invokes tools, and coordinates execut…
DeepSearch-World: Self-Distillation for Deep Search Agents in a Verifiable Environment
Xinyu Geng, Xuanhua He, Sixiang Chen +7
Training tool-use agents to improve from their own experience remains challenging, as supervised fine-tuning relies on fixed teacher-distilled trajectories, while sparse-reward rei…
Hierarchical Sparse Attention Done Right: Toward Infinite Context Modeling
Xiang Hu, Xinyu Wei, Hao Gu +10
Scaling modern large language models (LLMs) to long contexts is limited by the quadratic computation cost, and poor length extrapolation of dense attention. Chunk-wise sparse atten…
Learning to Build the Environment: Self-Evolving Reasoning RL via Verifiable Environment Synthesis
Yucheng Shi, Zhenwen Liang, Kishan Panaganti +3
We pursue a vision for self-improving language models in which the model does not merely generate problems or traces to imitate, but constructs the environments that train it. In z…
DeltaRubric: Generative Multimodal Reward Modeling via Joint Planning and Verification
Rui Liu, Dian Yu, Zhenwen Liang +6
Aligning Multimodal Large Language Models (MLLMs) requires reliable reward models, yet existing single-step evaluators can suffer from lazy judging, exploiting language priors over…
Reinforcing Multimodal Reasoning Against Visual Degradation
Rui Liu, Dian Yu, Haolin Liu +6
Reinforcement Learning has significantly advanced the reasoning capabilities of Multimodal Large Language Models (MLLMs), yet the resulting policies remain brittle against real-wor…
Measure Twice, Click Once: Co-evolving Proposer and Visual Critic via Reinforcement Learning for GUI Grounding
Wenkai Wang, Xiyun Li, Hongcan Guo +5
Graphical User Interface (GUI) grounding requires mapping natural language instructions to precise pixel coordinates. However, due to visually homogeneous elements and dense layout…
Locas: Your Models are Principled Initializers of Locally-Supported Parametric Memories
Sidi Lu, Zhenwen Liang, Dongyang Ma +3
In this paper, we aim to bridge test-time-training with a new type of parametric memory that can be flexibly offloaded from or merged into model parameters. We present Locas, a Loc…
Verified Critical Step Optimization for LLM Agents
Mukai Li, Qingcheng Zeng, Tianqing Fang +5
As large language model agents tackle increasingly complex long-horizon tasks, effective post-training becomes critical. Prior work faces fundamental challenges: outcome-only rewar…
Group Distributionally Robust Optimization-Driven Reinforcement Learning for LLM Reasoning
Kishan Panaganti, Zhenwen Liang, Wenhao Yu +2
Recent progress in Large Language Model (LLM) reasoning is increasingly driven by the refinement of post-training loss functions and alignment strategies. However, standard Reinfor…
Save the Good Prefix: Precise Error Penalization via Process-Supervised RL to Enhance LLM Reasoning
Haolin Liu, Dian Yu, Sidi Lu +6
Reinforcement learning (RL) has emerged as a powerful framework for improving the reasoning capabilities of large language models (LLMs). However, most existing RL approaches rely…
Inference-Time Scaling of Verification: Self-Evolving Deep Research Agents via Test-Time Rubric-Guided Verification
Yuxuan Wan, Tianqing Fang, Zaitang Li +5
Recent advances in Deep Research Agents (DRAs) are transforming automated knowledge discovery and problem-solving. While the majority of existing efforts focus on enhancing policy…
Stable and Efficient Single-Rollout RL for Multimodal Reasoning
Rui Liu, Dian Yu, Lei Ke +6
Reinforcement Learning with Verifiable Rewards (RLVR) has become a key paradigm to improve the reasoning capabilities of Multimodal Large Language Models (MLLMs). However, prevalen…
Can LLMs Guide Their Own Exploration? Gradient-Guided Reinforcement Learning for LLM Reasoning
Zhenwen Liang, Sidi Lu, Wenhao Yu +4
Reinforcement learning has become essential for strengthening the reasoning abilities of large language models, yet current exploration mechanisms remain fundamentally misaligned w…
Guided Self-Evolving LLMs with Minimal Human Supervision
Wenhao Yu, Zhenwen Liang, Chengsong Huang +4
AI self-evolution has long been envisioned as a path toward superintelligence, where models autonomously acquire, refine, and internalize knowledge from their own learning experien…