23 papers
RM-OPD: On-Policy Distillation for Generative Reward Models
Qinglin Su, Licheng Pan, Zhen Fang +5
Reinforcement Learning on Mixed-domain preference data (Mix-RL) has become the mainstream approach for training general-purpose Generative Reward Models (GRMs). However, Mix-RL suf…
ExeWorldBench: Benchmarking Code-Based Executable 3D World Modeling
Xinge Peng, Zhen Fang, Ruibing Wang +8
Coding agents are increasingly evolving beyond reproducing what a world looks like toward constructing executable worlds that capture how a world works. This shift calls for execut…
GraphForge: Training Working Agents with Graph-Anchored Workspace Synthesis
Qisheng Su, Hanchen Wang, Guanru Zhu +11
Working agents need to read diverse files, coordinate tools, and produce deliverables. Training such agents requires tasks built on many real files with verifiable results, but few…
Toolcompass: Guiding Tool Trialing, Not Suppressing It
Junlin Fang, Chong Zhang, Do Nguyen-Thanh +3
Large language model (LLM) agents must generalize from tools seen during training to unseen tools at deployment. A key challenge is tool trialing, i.e., excessive trials waste the…
FACET: Preserving Source Intent and Executable State in Terminal Task Synthesis
Kou Shi, Zun Wang, Qisheng Su +10
Training terminal agents requires scalable executable supervision, yet synthesizing high-quality terminal tasks remains challenging. Each task couples an instruction, an initialize…
VideoCoCo: Code-as-CoT for Physically-Consistent Video Generation via an Agentic Dual-Engine System
Haodong Li, Tianfei Ren, Xiaoxiao Ma +25
Text-to-video models have achieved remarkable visual quality, yet they still struggle to generate physically consistent dynamics because the temporal evolution of a scene must be i…
Video-DeepResearch: Towards the Next-Generation Multimodal Deepresearch Agent
Zhen Fang, Yu Zeng, Wenxuan Huang +17
We introduce Video-DeepResearch (Video-DR), extending multimodal agents from static images to continuous video streams, a setting that demands dense spatiotemporal grounding couple…
Reasoning Denoiser: Denoising Reasoning Traces for Hallucination Detection in Large Reasoning Models
Junlin Fang, Do Nguyen-Thanh, Xiaogang Xu +2
Large reasoning models (LRMs) generate long reasoning traces before producing final answers. While these traces may contain useful signals for hallucination detection, harnessing t…
ACC: Compiling Agent Trajectories for Long-Context Training
Qisheng Su, Zhen Fang, Shiting Huang +8
Recent development of agents has renewed demand for long-context reasoning capacity of LLMs. However, training LLMs for this capacity requires costly long-document curation or heur…
SaaSBench: Exploring the Boundaries of Coding Agents in Long-Horizon Enterprise SaaS Engineering
Qingnan Ren, Shun Zou, Shiting Huang +11
As autonomous coding agents become capable of handling increasingly long-horizon tasks, they have gradually demonstrated the potential to complete end-to-end software development.…
VideoSeeker: Incentivizing Instance-level Video Understanding via Native Agentic Tool Invocation
Yiming Zhao, Yu Zeng, Wenxuan Huang +11
Large Vision-Language Models (LVLMs) have shown significant progress in video understanding, yet they face substantial challenges in tasks requiring precise spatiotemporal localiza…
AsyncTool: Evaluating the Asynchronous Function Calling Capability under Multi-Task Scenarios
Kou Shi, Ziao Zhang, Shiting Huang +7
Large language model (LLM)-based agents have shown strong capabilities in using external tools to solve complex tasks. However, existing evaluations often overlook the temporal dim…
Beyond Accuracy: Unveiling Inefficiency Patterns in Tool-Integrated Reasoning
Qisheng Su, Shiting Huang, Zhen Fang +3
In real-world Tool-Integrated Reasoning (TIR) scenarios, where LLMs interleave reasoning with external tool calls, a major source of inefficiency is that the toolcalls create pause…
Vision-DeepResearch Benchmark: Rethinking Visual and Textual Search for Multimodal Large Language Models
Yu Zeng, Wenxuan Huang, Zhen Fang +14
Multimodal Large Language Models (MLLMs) have advanced VQA and now support Vision-DeepResearch systems that use search engines for complex visual-textual fact-finding. However, eva…
Internalizing Meta-Experience into Memory for Guided Reinforcement Learning in Large Language Models
Shiting Huang, Zecheng Li, Yu Zeng +7
Reinforcement Learning with Verifiable Rewards (RLVR) has emerged as an effective approach for enhancing the reasoning capabilities of Large Language Models (LLMs). Despite its eff…
ADORA: Training Reasoning Models with Dynamic Advantage Estimation on Reinforcement Learning
Qingnan Ren, Shiting Huang, Zhen Fang +4
Reinforcement learning has become a cornerstone technique for developing reasoning models in complex tasks, ranging from mathematical problem-solving to imaginary reasoning. The op…
Digital Metabolism: Decoupling Logic from Facts via Regenerative Unlearning -- Towards a Pure Neural Logic Core
Mengmeng Peng, Zhenyu Fang, He Sun
Large language models (LLMs) currently suffer from parameter entanglement, where general reasoning capabilities (logic) and specific factual knowledge (facts) exist in a superposit…
UniCorn: Towards Self-Improving Unified Multimodal Models through Self-Generated Supervision
Ruiyan Han, Zhen Fang, XinYu Sun +9
While Unified Multimodal Models (UMMs) have achieved remarkable success in cross-modal comprehension, a significant gap persists in their ability to leverage such internal knowledg…
Vision-DeepResearch: Incentivizing DeepResearch Capability in Multimodal Large Language Models
Wenxuan Huang, Yu Zeng, Qiuchen Wang +14
Multimodal large language models (MLLMs) have achieved remarkable success across a broad range of vision tasks. However, constrained by the capacity of their internal world knowled…
DualVLA: Building a Generalizable Embodied Agent via Partial Decoupling of Reasoning and Action
Zhen Fang, Zhuoyang Liu, Jiaming Liu +7
To build a generalizable Vision-Language-Action (VLA) model with strong reasoning ability, a common strategy is to first train a specialist VLA on robot demonstrations to acquire r…
CRITICTOOL: Evaluating Self-Critique Capabilities of Large Language Models in Tool-Calling Error Scenarios
Shiting Huang, Zhen Fang, Zehui Chen +6
The ability of large language models (LLMs) to utilize external tools has enabled them to tackle an increasingly diverse range of tasks. However, as the tasks become more complex a…
Provable Ordering and Continuity in Vision-Language Pretraining for Generalizable Embodied Agents
Zhizhen Zhang, Lei Zhu, Zhen Fang +2
Pre-training vision-language representations on human action videos has emerged as a promising approach to reduce reliance on large-scale expert demonstrations for training embodie…
Towards Control-Centric Representations in Reinforcement Learning from Images
Chen Liu, Hongyu Zang, Xin Li +5
Image-based Reinforcement Learning is a practical yet challenging task. A major hurdle lies in extracting control-centric representations while disregarding irrelevant information.…