Leveraging reinforcement learning, MPAR$^2$ improves perception performance on CAFE from 31.74% to 63.51% and effectively mitigates perception decay, concurrently enhancing reasoning capabilities to achieve a significant 74.59% accuracy on the MMAU benchmark.
RRC introduces two complementary strategies: self-competitive ranking, which exploits comparisons among sampled responses, and anchor-guided ranking, which enables scalable ranking-based reward construction with a small set of reference responses.
Chenglong Wang, Ziming Zhu, Yifu Huo et al.· 2 citations
Environmental Feedback-based Credit Assignment (EFCA), a multi-timescale credit assignment approach for long-horizon agentic RL that complements the long-term outcome signal with two environment-grounded process signals: a short-term feedback signal that captures the immediate effect of the current action and a medium-...
Yifu Huo, Shunjie Xing, Chenglong Wang et al.· 1 citation
Flow Continuous Trajectory Supervision (FlowCTS), which matches subsequent student and reference trajectories initialized from the same student-visited state to derive a temporally weighted velocity-matching upper bound and discretize it into practical objectives parameterized by the number of supervision steps.
Kaiyang Ye, Yuan Ge, Junxia Zhang et al.· arXiv.org· 0 citations
ToFu is presented, an agentic harness for researchers that reads your codebase, edits files, runs commands, and integrates with your development tools and provides a white-box agentic harness that allows researchers to inspect, modify, and evaluate its orchestration logic, tool-use behavior, and harness design.
Junhao Ruan, Yuan Ge, Bei Li et al.· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.
