WorkRover: Fudan Unveils Recursive Environment Evolution for Cowork Agents, Surging JobBench by 26.27%
Training long-horizon cowork agents via reinforcement learning requires scalable, iteratively refined interactive environments. Researchers from Fudan University NLP Lab introduced WorkRover, a framework enabling recursive environment evolution. Featuring Factorial Compatibility Replay (FCR) to maintain alignment across work contexts, tasks, and verifiers, and Verifier-to-Turn Credit Assignment (VTCA) to route progress signals to causal prerequisite turns, WorkRover-35B-A3B outperforms baseline checkpoints by 26.27% on JobBench, 29.32% on OfficeQA Pro, and 413.25 Elo on GDPval v2.
Key Takeaways
- Recursive environment evolution: empowers construction agents to iteratively adapt work contexts, tasks, and verifiers from workspace execution feedback
- FCR and VTCA algorithms: Factorial Compatibility Replay selects minimal valid repairs; Verifier-to-Turn Credit Assignment credits causal prerequisite turns
- Massive empirical leaps: WorkRover-35B-A3B achieves +26.27% on JobBench, +29.32% on OfficeQA Pro, and +413.25 Elo on GDPval v2
Linked Models & Agents:
⚡ WorkRover-35B-A3B⚡ Qwen3.5-35B-A3B🤖 Cowork Agents🤖 Office Automation Agents🌐 Fudan University🌐 Fudan NLP
