analiza aiqbench
This research addresses the silent failure modes of deep learning in autonomous driving by reframing out-of-distribution monitoring as a quickest changepoint detection problem. By modeling prediction error evolution with Hidden Markov Models, the authors establish a method that offers provable guarantees on detection delay and false-alarm rates without requiring prior knowledge of post-change distributions. The approach demonstrates robustness across real-world datasets where standard uncertainty metrics fail to rise despite collapsing accuracy.
Latent Dynamics-Aware OOD Monitoring for Trajectory Prediction with Provable Guarantees
arXiv:2603.14603v2 Announce Type: replace Abstract: In safety-critical Cyber-Physical Systems (CPS), trajectory prediction guides downstream planning and control. Deep learning models forecast well on validation data, but their reliability drops in out-of-distribution (OOD) scenarios driven by environmental uncertainty or rare traffic behaviors [1, 2]. Such failures are often silent: forecasts stay spatially plausible while accuracy collapses, and reported uncertainty does not rise [3]. Detection is hard because traffic conditions and interaction patterns keep evolving, yet the safety-critical nature of autonomous driving (AD) demands formal guarantees on detection delay and false-alarm rate. Following [4], we reframe OOD monitoring as quickest changepoint detection (QCD), a principled statistical framework with well-established theory. We find that the evolution of prediction errors on in-distribution (ID) data is well modeled by a Hidden Markov Model (HMM). Building on this, we extend a recent cumulative Maximum Mean Discrepancy approach to our setting. The method needs no detailed prior knowledge of the post-change distribution, yet admits provable delay and false-alarm guarantees. On three real-world driving datasets, it reduces detection delay while staying robust to heavy-tailed distributions and unknown post-change conditions.
Abstract:In safety-critical Cyber-Physical Systems (CPS), trajectory prediction guides downstream planning and control. Deep learning models forecast well on validation data, but their reliability drops in out-of-distribution (OOD) scenarios driven by environmental uncertainty or rare traffic behaviors [1, 2]. Such failures are often silent: forecasts stay spatially plausible while accuracy collapses, and reported uncertainty does not rise [3]. Detection is hard because traffic conditions and interaction patterns keep evolving, yet the safety-critical nature of autonomous driving (AD) demands formal guarantees on detection delay and false-alarm rate. Following [4], we reframe OOD monitoring as quickest changepoint detection (QCD), a principled statistical framework with well-established theory. We find that the evolution of prediction errors on in-distribution (ID) data is well modeled by a Hidden Markov Model (HMM). Building on this, we extend a recent cumulative Maximum Mean Discrepancy approach to our setting. The method needs no detailed prior knowledge of the post-change distribution, yet admits provable delay and false-alarm guarantees. On three real-world driving datasets, it reduces detection delay while staying robust to heavy-tailed distributions and unknown post-change conditions.
Robotics (cs.RO)
arXiv:2603.14603 [cs.RO]
(or arXiv:2603.14603v2 [cs.RO] for this version)
https://doi.org/10.48550/arXiv.2603.14603 Focus to learn more arXiv-issued DOI via DataCite
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Text integral sub licența deschisă a publicației · 267 cuvinte
Subiecte conexe
STELLAR: Spatio-Temporal Environmental Learning with Latent Alignment and Refinement for Long-Tailed Species Distribution Modeling
arXiv · 9 iun. 2026
PIRAMID: Progress and Plans
LessWrong · 27 iul. 2026
GlucoFM: Foundation model for continuous glucose monitoring
Google Research · 26 aug. 2026
Research Note: Filtering Subversion-Relevant Information From Pretraining Data Is Feasible
LessWrong · 5 oct. 2026