20 papers
KPHarness: Co-Evolving Knowledge and Probing for Long-Horizon Gaming Harnesses
Zhikang Chen, Yueqing Sun, Sen Cui +6
A failed gaming trajectory reveals that an agent needs intervention, but not necessarily how its execution harness should change. Effective harness adaptation therefore requires ga…
Structured Latent Dynamics for Long-Horizon Embodied World Models
Kevin Chen, Yu-Wei Zhan, Haoran Liang +2
Reliable long-horizon prediction remains a central challenge for embodied world models. Although recent generative models can synthesize visually plausible action-conditioned futur…
RSIFlow: Adaptive Self-Improvement Workflows with an Evolving Meta-Agent
Huan Zhang, Mingju Chen, Dongxu Zhou +4
Large language model (LLM) agents have demonstrated strong capabilities in complex tasks, including code generation, tool use, and multi-hop question answering. Existing self-impro…
RoboCoach: World Models as Active Coaches for Compositional Robot Skills
Jiajun Liu, Yifan Chen, Yichao Liu +7
Long-horizon robot manipulation reuses skills across many task compositions, but improving these compositions with additional end-to-end demonstrations is costly. A practical self-…
Unifying Distributional Training for One-Step Visual Generation
Chi Zhang, Shi Haoyang, Haoyang Shi +10
Distributional training provides collective supervision for one-step visual generation by matching real and generated features in frozen representation spaces. We introduce a unifi…
From Scores to Samples: Elastic Forcing for Autoregressive Video Generation
Chi Zhang, Yueyi Liu, Shi Haoyang +6
Few-step autoregressive video generation commonly relies on Distribution Matching Distillation (DMD), requiring a bidirectional diffusion teacher and an online fake-score model. We…
HamiFormer: Dual-Expert Diffusion Fields with Affine Symplectic Maps
Haoxiang Huang, Xiang Liu, Shuwei Wang +2
Predicting smooth dynamics and collisions requires modeling continuous evolution and abrupt state changes. We introduce HamiFormer, a dual-expert diffusion field combining whole-wi…
LPA-CWM: A Learned Physical Adjudicator for Motion Reasoning with Counterfactual World Models
Kunwei Wu, Xiang Liu, Guocai Yao +5
Counterfactual world models (CWM) extract motion from pretrained video predictors by comparing factual and intervened predictions, but uniform aggregation weights responses equally…
MOSH-WM: Mask-Grounded Soft-Hamiltonian Dynamics for Object-Centric World Models
Zhekai Wang, Haoxiang Huang, Xiang Liu +6
Object-centric world models forecast future videos by evolving a set of entity slots, but the variables receiving dynamics supervision are often unconstrained visual features. We i…
TRCA: Transition-wise Rubric Credit Assignment for Long-horizon LLM Agents
Huan Zhang, Mingju Chen, Dongxu Zhou +5
Long-horizon large language model (LLM) agents are typically optimized with sparse terminal outcomes, making fine-grained credit assignment across multi-step interactions difficult…
VERDI:Verified Contradictions for Continual World-Model Optimization
Junyu Wu, Shiqin Nie, Youyi Kou +10
Foundation world models have made remarkable progress in planning, simulation, and embodied intelligence. However, optimizing a pretrained world model toward a user-specified objec…
ElasticTTT: Prior-Preserving Test-Time Tuning for Video Editing
Yueyi Liu, Chi Zhang, Sen Cui +1
Test-Time Tuning (TTT) on pretrained diffusion models has emerged as a powerful paradigm for video editing. However, there exists a foundational mismatch between the distribution-m…
Physically Native World Models: A Hamiltonian Perspective on Generative World Modeling
Sen Cui, Jingheng Ma
World models have recently re-emerged as a central paradigm for embodied intelligence, robotics, autonomous driving, and model-based reinforcement learning. However, current world…
MetaForge: A Self-Evolving Multimodal Agent that Retrieves, Adapts, and Forges Tools On Demand
Shouang Wei, Houcheng Min, Xinpeng Dong +8
Multimodal agents have achieved notable progress on complex reasoning tasks through tool use, yet remain limited by two issues: statically predefined tool inventories fail to gener…
Guideline-grounded retrieval-augmented generation for ophthalmic clinical decision support
Shuying Chen, Sen Cui, Zhong Cao
In this work, we propose Oph-Guid-RAG, a multimodal visual RAG system for ophthalmology clinical question answering and decision support. We treat each guideline page as an indepen…
FATE: Closed-Loop Feasibility-Aware Task Generation with Active Repair for Physically Grounded Robotic Curricula
Bingchuan Wei, Bingqi Huang, Jingheng Ma +2
Recent breakthroughs in generative simulation have harnessed Large Language Models (LLMs) to generate diverse robotic task curricula, yet these open-loop paradigms frequently produ…
Reversible Diffusion Decoding for Diffusion Language Models
Xinyun Wang, Min Zhang, Sen Cui +4
Diffusion language models enable parallel token generation through block-wise decoding, but their irreversible commitments can lead to stagnation, where the reverse diffusion proce…
Merging without Forgetting: Continual Fusion of Task-Specific Models via Optimal Transport
Zecheng Pan, Zhikang Chen, Ding Li +9
Merging models fine-tuned for different tasks into a single unified model has become an increasingly important direction for building versatile, efficient multi-task systems. Exist…
Think Consistently, Reason Efficiently: Energy-Based Calibration for Implicit Chain-of-Thought
Zhikang Chen, Sen Cui, Deheng Ye +3
Large Language Models (LLMs) have demonstrated strong reasoning capabilities through \emph{Chain-of-Thought} (CoT) prompting, which enables step-by-step intermediate reasoning. How…
CoDoL: Conditional Domain Prompt Learning for Out-of-Distribution Generalization
Min Zhang, Yuyin Wang, Zhongxiang Dai +4
Recent advances in pre-training vision-language models (VLMs), e.g., contrastive language-image pre-training (CLIP) methods, have shown great potential in learning out-of-distribut…