aiqbench analysis
FactoryNet addresses a critical gap in industrial AI by standardizing heterogeneous actuator data into a unified control-theoretic schema, enabling cross-manufacturer transfer learning that was previously impossible with vendor-specific logs. The study reveals that while simulation can capture stable task structures, fault signatures remain highly diverse and manufacturer-specific, requiring specialized calibration of motor current signals rather than generic feature alignment to bridge the sim-to-real gap. This work establishes a foundational dataset for training robust industrial foundation models capable of generalizing across different robotic embodiments.
FactoryNet: A Large-Scale Dataset toward Industrial Time-Series Foundation Models
arXiv:2605.09081v5 Announce Type: replace Abstract: Industrial machines are actuated systems, yet most open time-series corpora record sensor outcomes alone; the few that do log commanded signals expose them in vendor-specific forms that differ from machine to machine, so commanded intent and measured response cannot be compared across embodiments. We introduce FactoryNet, a pretraining corpus for industrial time-series: 113M datapoints across 57k end-to-end task executions (15k real, 42k synthetic) on seven embodiments and four tasks, with 27 annotated anomaly types alongside healthy baselines and counterfactual pairs. Every signal in every source is mapped into a single control-theoretic schema, Setpoint, Effort, Feedback, Context (S-E-F-C), which expresses any actuated system in a common representational frame and makes commanded-versus-realized dynamics readable as explicit prediction residuals. Using the corpus, we find that a task-phase model trained on simulation, with no real phase labels, recovers roughly two thirds of the real-data ceiling on both robots, and that transfer holds within a robot family but rarely across manufacturers. Fault signatures behave differently: an uncalibrated simulator gives no detectable fault signal on either robot, and aligning the feature distributions does not repair it. What separation there is comes from Effort: motor current distinguishes faults where measured tool motion is close to chance, and calibrating the simulator to reproduce that one channel recovers a payload fault on the real robot. We read this as simulation capturing the stable structure of a task while fault behaviour stays diverse and fault-type specific. We release FactoryNet as a growing, multi-embodiment substrate for industrial foundation models.
Authors:Jonas Petersen, Pratik Jawahar, Riccardo Maggioni, Alexandre Huou, Karim Othman, Camilla Mazzoleni, Matteo Cozzi, Marek Gajewski, Davide Paloschi, Gian-Alessandro Lombardi, Matei C. Ignuta-Ciuncanu, Lior Porath, Federico Martelli, Philipp Christian Petersen
Abstract:Industrial machines are actuated systems, yet most open time-series corpora record sensor outcomes alone; the few that do log commanded signals expose them in vendor-specific forms that differ from machine to machine, so commanded intent and measured response cannot be compared across embodiments. We introduce FactoryNet, a pretraining corpus for industrial time-series: 113M datapoints across 57k end-to-end task executions (15k real, 42k synthetic) on seven embodiments and four tasks, with 27 annotated anomaly types alongside healthy baselines and counterfactual pairs. Every signal in every source is mapped into a single control-theoretic schema, Setpoint, Effort, Feedback, Context (S-E-F-C), which expresses any actuated system in a common representational frame and makes commanded-versus-realized dynamics readable as explicit prediction residuals. Using the corpus, we find that a task-phase model trained on simulation, with no real phase labels, recovers roughly two thirds of the real-data ceiling on both robots, and that transfer holds within a robot family but rarely across manufacturers. Fault signatures behave differently: an uncalibrated simulator gives no detectable fault signal on either robot, and aligning the feature distributions does not repair it. What separation there is comes from Effort: motor current distinguishes faults where measured tool motion is close to chance, and calibrating the simulator to reproduce that one channel recovers a payload fault on the real robot. We read this as simulation capturing the stable structure of a task while fault behaviour stays diverse and fault-type specific. We release FactoryNet as a growing, multi-embodiment substrate for industrial foundation models.
Accepted at AI4Physics and FMSD, ICML 2026. Code: this https URL
Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
arXiv:2605.09081 [cs.LG]
(or arXiv:2605.09081v5 [cs.LG] for this version)
Full text under the publisher's open licence · 303 words
Related coverage
News
Quantum Computing Report · Jun 7, 2026
FactoryNet: A Large-Scale Dataset toward Industrial Time-Series Foundation Models
arXiv · Jun 5, 2026
Dario Amodei, The Physicist Who Left OpenAI to Build Anthropic
Quantum Zeitgeist · Aug 30, 2026
XtalPi Holdings Limited: XtalPi Holdings Announces 2026 Interim Results
finanznachrichten.de · Aug 19, 2026