Safe Reinforcement Learning With Model Uncertainty Estimates
Open accessThe DOI matches its Crossref record; 43 of the 43 references it cites were found in a registry or an index.
Abstract
Many current autonomous systems are being designed with a strong reliance on black box predictions from deep neural networks (DNNs). However, DNNs tend to be overconfident in predictions on unseen data and can give unpredictable results for far-from-distribution test data. The importance of predictions that are robust to this distributional shift is evident for safety-critical applications, such as collision avoidance around pedestrians. Measures of model uncertainty can be used to identify unseen data, but the state-of-the-art extraction methods such as Bayesian neural networks are mostly intractable to compute. This paper uses MC-Dropout and Bootstrapping to give computationally tractable and parallelizable uncertainty estimates. The methods are embedded in a Safe Reinforcement Learning framework to form uncertainty-aware navigation around pedestrians. The result is a collision avoidance policy that knows what it does not know and cautiously avoids pedestrians that exhibit unseen behavior. The policy is demonstrated in simulation to be more robust to novel observations and take safer actions than an uncertainty-unaware baseline.
Overview
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
Citations
Cites
Hochreiter, S., & Schmidhuber, J. (1997) · Neural Computation · doi:10.1162/neco.1997.9.8.1735 · Cited by 102,318 · cited 10×
“A set of LSTM networks (ensemble) estimates the probability P ( coll | u t − l : t + h , o t − l : t ) that a motion primitive u t : t + h would lead to a collision in the next h time steps, given the history of observations o t − l : t and past actions u t − l : t .”
InfluentialAn Introduction to the Bootstrap
Efron, B., & Tibshirani, R. J. (1994) · doi:10.1201/9780429246593 · Cited by 40,500
(2014) · Cited by 33,419
An Introduction to the Bootstrap
Efron, B., & Tibshirani, R. J. (1993) · doi:10.1007/978-1-4899-4541-9 · Cited by 21,230
Social force model for pedestrian dynamics
Helbing, D., & Molnár, P. (1995) · Physical Review E · doi:10.1103/physreve.51.4282 · Cited by 6,845 · cited 1×
“Humans achieve mostly collisionfree navigation by understanding the hidden intentions of other pedestrians and vehicles and interacting with them [7], [8].”
BackgroundDropout as a bayesian approximation: Representing model uncertainty in deep learning
Gal, Y., & Ghahramani, Z. (2015) · arXiv · doi:10.48550/arxiv.1506.02142 · Cited by 6,471
Bayesian Learning for Neural Networks
Neal, R. M. (1996) · Lecture Notes in Statistics · doi:10.1007/978-1-4612-0745-0 · Cited by 4,276 · cited 2×
“Mostly, Bayesian neural networks are used to extract model uncertainty but require a significant restructuring of the network architecture [20].”
BackgroundMethodSimple and Scalable Predictive Uncertainty Estimation using Deep Ensembles
Lakshminarayanan, B., Pritzel, A., & Blundell, C. (2016) · arXiv · doi:10.48550/arxiv.1612.01474 · Cited by 3,666 · cited 3×
“extensive computational cost and have a sample-dependent accuracy [2], [20], [24].”
BackgroundMethodSocial LSTM: Human Trajectory Prediction in Crowded Spaces
Alahi, A., Goel, K., Ramanathan, V., Robicquet, A., Fei-Fei, L., & Savarese, S. (2016) · 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2016.110 · Cited by 3,483 · cited 1×
“LSTM networks are chosen for the dynamic obstacle avoidance, because they are the state-of-the-art model in predicting pedestrian paths by understanding the hidden temporal intentions of pedestrians best [42], [43].”
BackgroundWhat Uncertainties Do We Need in Bayesian Deep Learning for Computer Vision?
Kendall, A., & Gal, Y. (2017) · arXiv · doi:10.48550/arxiv.1703.04977 · Cited by 3,088
Cited by
A review of uncertainty quantification in deep learning: Techniques, applications and challenges
Abdar, M., Pourpanah, F., Hussain, S., Rezazadegan, D., Liu, L., Ghavamzadeh, M., Fieguth, P., Cao, X., Khosravi, A., Acharya, U. R., et al. (2021) · Information Fusion · doi:10.1016/j.inffus.2021.05.008 · Cited by 3,111
“[184] SRL UA navigation Ensemble of MC dropout (EMCD) and Bootstrapping ×”
A survey of uncertainty in deep neural networks
Gawlikowski, J., Tassi, C. R. N., Ali, M., Lee, J., Humt, M., Feng, J., Kruspe, A., Triebel, R., Jung, P., Roscher, R., et al. (2023) · Artificial Intelligence Review · doi:10.1007/s10462-023-10562-9 · Cited by 1,406
Safe Learning in Robotics: From Learning-Based Control to Safe Reinforcement Learning
Brunke, L., Greeff, M., Hall, A. W., Yuan, Z., Zhou, S., Panerati, J., & Schoellig, A. P. (2022) · Annual Review of Control, Robotics, and Autonomous Systems · doi:10.1146/annurev-control-042920-020211 · Cited by 769
Explainability in deep reinforcement learning
Heuillet, A., Couthouis, F., & Díaz-Rodríguez, N. (2021) · Knowledge-Based Systems · doi:10.1016/j.knosys.2020.106685 · Cited by 319
A Review of Safe Reinforcement Learning: Methods, Theories, and Applications
Gu, S., Yang, L., Du, Y., Chen, G., Walter, F., Wang, J., & Knoll, A. (2024) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2024.3457538 · Cited by 315
Ganesh, A. H., & Xu, B. (2022) · Renewable and Sustainable Energy Reviews · doi:10.1016/j.rser.2021.111833 · Cited by 285
Safe Reinforcement Learning With Stability Guarantee for Motion Planning of Autonomous Vehicles
Zhang, L., Zhang, R., Wu, T., Weng, R., Han, M., & Zhao, Y. (2021) · IEEE Transactions on Neural Networks and Learning Systems · doi:10.1109/tnnls.2021.3084685 · Cited by 165
Safe Reinforcement Learning for Autonomous Lane Changing Using Set-Based Prediction
Krasowski, H., Wang, X., & Althoff, M. (2020) · 2020 IEEE 23rd International Conference on Intelligent Transportation Systems (ITSC) · doi:10.1109/itsc45102.2020.9294259 · Cited by 76
“Various approaches have been proposed to increase the safety of RL methods by modifying the optimality criterion [1], [2] or by verifying the exploration processes with external guidance [3]–[10].”
MethodBlind spot detection for safe sim-to-real transfer
Ramakrishnan, R., Kamar, E., Dey, D., Horvitz, E., & Shah, J. (2020) · Journal of Artificial Intelligence Research · doi:10.1613/jair.1.11436 · Cited by 21
“Many other works have developed approaches for estimating model uncertainty using various techniques (L¨utjens, Everett, & How, 2018; McAllister, Gal, Kendall, Van Der Wilk, Shah, Cipolla, & Weller, 2017; Malinin & Gales, 2018).”
Method
Metadata
- Authors
- Lutjens, B., Everett, M., & How, J. P.
- Venue
- 2019 International Conference on Robotics and Automation (ICRA)
- Publisher
- IEEE
- Published
- 2019-05
- Pages
- 8662-8668
- ISSN
- 1050-4729 (ISSN-L)
- OpenAlex
- W2898045695
- Cited by
- 171
- References
- 43
- FWCI
- 8.01
- Type
- conference paper
- Language
- en
- Open access
- green
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.