Reproduce “Runtime Monitoring of Perception-Based Autonomous Systems via Embedding Temporal Logic” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Parv Kapoor, Abigail Hammer, Ashish Kapoor, Karen Leung, Eunsuk Kang
ETL monitors robot behavior by comparing scene embeddings with reference images and composing thresholded similarity predicates into temporal rules. Across five DROID tasks, it achieved mean phase F1 of 0.813 and agreement of 0.940, versus 0.390 and 0.567 for a prompted monitor, and detected milestone order correctly in four tasks. It scored lower on the four-phase chain; these are monitoring, not task-success, metrics.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.