activity
20232026
most citedProgressive Bird's Eye View Perception for Safety-Critical Autonomous Driving: A Comprehensive Survey

1 citations · 1 across the 11 of their papers we have counts for

collaborators
Showing cs.LGShow all

6 papers · 1 filter

cs.LG2026

OpenWebRL: Demystifying Online Multi-turn Reinforcement Learning for Visual Web Agents

Rui Yang, Qianhui Wu, Yuxi Chen +7

Building capable visual web agents requires long-horizon reasoning, precise grounding, and robust interaction with dynamic real-world websites. Despite rapid progress, the stronges…

cs.LG2026

AsyncWebRL: Efficient Asynchronous Reinforcement Learning for Multi-Step Visual Web Agents

Hao Bai, Rui Yang, Chenlu Ye +3

Training vision-language web agents with multi-step RL is compute-intensive, with two dominant forms of inefficiency: idle GPUs in synchronous RL, and trajectories that use more st…

cs.LG2026

InT: Self-Proposed Interventions Enable Credit Assignment in LLM Reasoning

Matthew Y. R. Yang, Hao Bai, Ian Wu +3

Outcome-reward reinforcement learning (RL) has proven effective at improving the reasoning capabilities of large language models (LLMs). However, standard RL assigns credit only at…

cs.LG2026

WebGym: Scaling Training Environments for Visual Web Agents with Realistic Tasks

Hao Bai, Alexey Taymanov, Tong Zhang +2

We present WebGym, the largest-to-date open-source environment for training realistic visual web agents. Real websites are non-stationary and diverse, making artificial or small-sc…

cs.LG2025

Thinking vs. Doing: Agents that Reason by Scaling Test-Time Interaction

Junhong Shen, Hao Bai, Lunjun Zhang +8

The current paradigm of test-time scaling relies on generating long reasoning traces ("thinking" more) before producing a response. In agent problems that require interaction, this…

cs.LG2025

Digi-Q: Learning Q-Value Functions for Training Device-Control Agents

Hao Bai, Yifei Zhou, Li Erran Li +2

While a number of existing approaches for building foundation model agents rely on prompting or fine-tuning with human demonstrations, it is not sufficient in dynamic environments…