LGWM: A World Model That Verifies GUI Agent Transitions and Detects Hijacked Screens
Jiaming Zhang · hf · 2026-10-22
NTU, Institute of Science Tokyo, and Sony AI present LGWM, a decoder-free action-conditioned world model that predicts next-screen representations to verify GUI transitions. Trained on 1.85M real GUI transitions without semantic annotation, it reduces verification to a vector comparison and flags harmful mismatches. The paper argues safety is a property of transitions, not screens, and introduces RSWT-BENCH, where screen-only detectors perform at chance.
More from coding & agent
- OpenLore gives every coding agent the same memory with one folder of Markdown — Arindam_1729 · 2026-10-22
- Walking bento box built with Claude Code + Blender Python, no MCP needed — sidahuj · 2026-10-22
- Claude Code Autonomously Generated 2,000+ Krea2 Images Overnight — reynadsaltynuts · 2026-10-22
- Indie dev building game #4 with AI starts with a central game asset management hub — ezshine · 2026-10-22
- Diffy: a cross-platform desktop code review app with a custom semantic diff engine — seatedro · 2026-10-22
- Open-Source Garmin MCP Connector Plugs Workout Data Into AI Agents — Vast-Grapefruits · 2026-10-22