DeskForge: Dense Supervision from Desktop Environments for Computer-Use Agents
arXiv:2610.02320v1 Announce Type: new Abstract: Computer-use agents need to reliably ground action targets in complex desktop scenes, where multiple applications, overlapping windows, and visually similar controls compete for attention. Existing training data rarely pair such scenes with dense anno…
Read the full story at arXiv cs.CV ↗
Timeline · 1 report
- 2026-10-05 04:00 · arXiv cs.CV
DeskForge: Dense Supervision from Desktop Environments for Computer-Use Agents