Tag
This paper presents a Product-Aware Autoencoder for robust anomaly detection in multi-product cyber-physical systems, addressing the blind spot of global models. It demonstrates improved detection accuracy over baseline in stress tests.
This paper presents a retrospective analysis of the CODS 2025 AssetOpsBench challenge, evaluating multi-agent AI systems on industrial tasks. It highlights discrepancies between public and hidden leaderboards and offers diagnostics for future agentic benchmarks.