Discriminative World Models for Web Agents
技术动态AI 50
来源：arXiv cs.AI发布时间待核实
arXiv:2609.02885v1 Announce Type: new Abstract: Recent web agents use world models for test-time action selection by sampling candidate actions, predicting the resulting web states, and ranking them with a ranker model or a Process Reward Model (PRM). These world models are typically trained via supervised next-state prediction to generate fixed representations like HTML or AXTree snapshots.