As video models enable longer and more realistic interactive experiences, maintaining a world consistent with its history becomes essential. This requires both preserving previously observed content and ensuring coherent world evolution over time. World memory is a critical capability to evaluate alongside visual fidel...
Nan-Yi Zhou, Jianxiong Gao, Qiuyu Wang et al.· 0 citations
Skill-augmented agents improve sample efficiency by distilling successful trajectories into reusable strategies. Yet most existing approaches remain text-centric, linearizing spatial layouts and action-state correspondences into language that loses critical geometric structure. Recent efforts have begun incorporating v...
Hong-Xing Li, Ding-Ming Li, Yi-Xin Li et al.· 0 citations
Existing spatial reasoning benchmarks mainly test spatial perception: reading off relations already visible in the input. Yet real-world spatial intelligence demands predictive spatial reasoning: constructing a scene from observations, anticipating how an intervention changes it, and reasoning about the unseen outcome....
Hong-Xing Li, Jin-Yue Su, Ding-Ming Li et al.· 0 citations
This work introduces WorldScope, which composes cross-view entity evidence and support-surface coverage into an image-subset evidence lattice, and proposes WorldFlow, which composes cross-view entity evidence and support-surface coverage into an image-subset evidence lattice.
Wei Wang, Wen-Qiao Zhang, Yu-Tong Lin et al.· 0 citations
This work proposes EvoSkill-GUI, a training-free framework in which each skill is a structured multi-file package containing retrieval metadata, executable plans, backup localization, failure-recovery rules, accessibility utilities, and failure cases.
Bo-Fan Chen, Bo-Xuan Zhang, Fei Tang et al.· 2 citations
BrowserForge is presented, a framework that generates web interaction data at scale by driving many browser sandboxes in parallel over the open web, and fine-tuning a compact multimodal model on this corpus raises its success rate on the live Online-Mind2Web from 25.66% to 33.33% and consistently improves step accuracy...
Fei Tang, Hua-Wen Shen, Zhiqiong Lu et al.· 0 citations
GSM8K-V is introduced, a benchmark transforming GSM8K into multi-image sequences with semantic equivalence preserved and error analysis shows that the primary bottleneck lies in Implicit Visual Inference Error (IVIE), where models fail to recover visual semantics that are implied rather than explicitly stated.
Fan Yuan, Yuchen Yan, Yi-Fan Jiang et al.· arXiv.org· 8 citations
This work presents EasySteer, a unified framework for high-performance, extensible LLM steering built on vLLM, and demonstrates its effectiveness in overthinking mitigation, hallucination reduction, and other key applications.
Haolei Xu, Xinyu Mei, Yuchen Yan et al.· arXiv.org· 14 citations
Test-Time Policy Optimization is proposed, an asymmetric objective that distills agreeing rollouts via OPSD and penalizes disagreeing rollouts with Grouped RL and Token-level selection further refines both branches: distillation down-weights already-converged positions, while RL penalizes only confident errors.
Ao-Han Wang, Zhengxi Lu, Jian-Ze Wang et al.· 2 citations
Results show that self-contamination is a trainable component of the LiC gap, and propose MAIGO, an on-policy self-distillation method that reduces this contamination using history-cleaned references from the model's own policy.
This work introduces PaperGym, a unified framework that turns each research paper into a complete training environment, and releases the pipeline, the 20,000-instance corpus PaperGym-20k, and the benchmarks PaperGym-Innov and PaperGym-Design.
Yu-Han Wang, Zhengxi Lu, Yuchen Yan et al.· 2 citations
EMPO achieves substantial gains over the base model and achieves competitive performance against strong baselines such as UI-TARS-7B and GPT-4o, demonstrating better generalization than prior single-turn RL approaches.
Zhengxi Lu, Jiabo Ye, Fei Tang et al.· Annual Meeting of the Associ...· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.
