Evaluation of autonomous systems under data distribution shifts
Not on the map. Why?
Abstract (arXiv)
We posit that data can only be safe to use up to a certain threshold of the data distribution shift, after which control must be relinquished by the autonomous system and operation halted or handed to a human operator. With the use of a computer vision toy example we demonstrate that network predictive accuracy is impacted by data distribution shifts and propose distance metrics between training and testing data to define safe operation limits within said shifts. We conclude that beyond an empirically obtained threshold of the data distribution shift, it is unreasonable to expect network predictive accuracy not to degrade
Key facts. Cited by 2 papers in the map (data through 7 Oct 2026). Builds on 13 of its 35 references that are in the map. First posted on arXiv 28 Jun 2024.
Builds on
13 of 35 references.
References · 13
- 2014Explaining and Harnessing Adversarial Examplescited by 7,287
- 2016A Baseline for Detecting Misclassified and Out-of-Distribution Examples in Neural Networkscited by 2,362
- 2018A Simple Unified Framework for Detecting Out-of-Distribution Samples and Adversarial Attackscited by 1,472
- 2017Enhancing The Reliability of Out-of-distribution Image Detection in Neural Networkscited by 1,325
- 2019Can You Trust Your Model's Uncertainty? Evaluating Predictive Uncertainty Under Dataset Shiftcited by 1,274
- 2016End to End Learning for Self-Driving Carscited by 1,000
- 2020CSI: Novelty Detection via Contrastive Learning on Distributionally Shifted Instancescited by 441
- 2020A Unifying Review of Deep and Shallow Anomaly Detectioncited by 423
- 2020Generalized ODIN: Detecting Out-of-distribution Image without Learning from Out-of-distribution Datacited by 408
- 2019Deep Learning for Anomaly Detection: A Surveycited by 357
- 2021MOS: Towards Scaling Out-of-distribution Detection for Large Semantic Spacecited by 233
- 2018Generative Probabilistic Novelty Detection with Adversarial Autoencoderscited by 99
1 more
Built on by
Citing papers per year: 2025: 1; 2026: 1.
2 papers.