ECHOES: Recursive History Compression for Long-Horizon GUI Agents
In brief
- ECHOES replaces screenshots with recursive textual states to cut inference costs.
- Online distillation guided by privileged data improves memory fidelity and task performance.
- The model delivers competitive results without storing visual history in production.
ECHOES introduces a method to compress visual history for GUI agents by replacing screenshots with recursively updated textual states. The model uses online distillation guided by privileged information to enhance memory fidelity and long-term task performance.
This summary gives you the essentials. The full article, its figures and quotes are on cs.AI updates on arXiv.org.
Read the original articleAI-generated summary of the source, checked by rules. Report an error