DRIVE: Modeling Skills at the Reasoning and Interaction Levels for Web Agents under Continual Learning
DRIVE framework boosts web agent performance by separating transferable reasoning from site-specific interactions.
Web agents require both high-level reasoning (for task decomposition) and low-level interactions (for page elements manipulation) to conduct different tasks. However, these knowledge types differ fundamentally: reasoning knowledge (e.g., booking a flight requires first searching ...
ArXiv CS.AI
Source
Generated insight by OnAirToday AI
The DRIVE framework addresses a fundamental bottleneck in agentic AI by decoupling abstract task logic from site-specific interaction mechanics. This modular approach significantly improves generalization, allowing agents to retain learned reasoning strategies while dynamically adapting to the unique interfaces of new web environments.
Comments
Keep comments relevant and respectful. No links, advertising, self-promotion, or spam.