Efficient statistical validation of machine learning systems for autonomous driving

Shi, Weijing; Alawieh, Mohamed Baker; Li, Xin; Yu, Huafeng; Aréchiga, Nikos; Tomatsu, Nobuyuki

doi:10.1145/2966986.2980077

Cited by 13 publications

(4 citation statements)

References 10 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…For degenerate metrics 𝑔 0 , non-zero tangent vectors can have zero length, since 𝑔 0 (𝑣, 𝑣) = 0. 5 Weijing Shi et al [26] state that a misclassification probability of 𝑝 ≈ 10 −12 would require a sample size of 𝑛 ≈ 10 13 .…”

Section: Then Pmentioning

confidence: 99%

Probabilistic Risk Assessment of an Obstacle Detection System for GoA 4 Freight Trains

Gleirscher,

Haxthausen,

Peleska

2023

Proceedings of the 9th ACM SIGPLAN International Workshop on Formal Techniques for Safety-Critical Systems

View full text Add to dashboard Cite

We propose a quantitative risk assessment approach for the design of an obstacle detection function for low-speed freight trains with grade of automation 4. In this five-step approach, starting with single detection channels and ending with a three-out-of-three model constructed of three independent dual-channel modules and a voter, we exemplify a probabilistic assessment, using a combination of statistical methods and parametric stochastic model checking. We illustrate that, under certain not unreasonable assumptions, the resulting hazard rate becomes acceptable for specific application settings. The statistical approach for assessing the residual risk of misclassifications in convolutional neural networks and conventional image processing software suggests that high confidence can be placed into the safety-critical obstacle detection function, even though its implementation involves realistic machine learning uncertainties.

show abstract

Section: Then Pmentioning

confidence: 99%

Probabilistic Risk Assessment of an Obstacle Detection System for GoA 4 Freight Trains

Gleirscher,

Haxthausen,

Peleska

2023

Proceedings of the 9th ACM SIGPLAN International Workshop on Formal Techniques for Safety-Critical Systems

View full text Add to dashboard Cite

show abstract

“…Reference [77] studies a use case of an object detection system that uses ML algorithms and provides an efficient approach to validate the system. For an image recognition algorithm to function properly in safety-critical automotive applications, a failure rate as small as 10 −12 should be achieved.…”

Section: B Challenges In Assessing Machine Learningmentioning

confidence: 99%

“…In this case, simulation approaches should be used, which in turn generates high costs if the model of the system is required to be accurate. To reduce validation cost in such a system, reference [77] proposes an approach of subset sampling in which the objective is to estimate the failure rate of an ML algorithm in an AV in terms of different probabilities. By experimenting within the use case of STOP sign detection,…”

Section: B Challenges In Assessing Machine Learningmentioning

confidence: 99%

Software Verification and Validation of Safe Autonomous Cars: A Systematic Literature Review

et al. 2021

View full text Add to dashboard Cite

“…Any visual perception system based on machine learning cannot be 100 [2]. Hence, the system may fail for a specific input pattern and accurately estimating its failure rate can be extremely time-consuming [3]. One can affirm in this case that the AI will not be necessary better than humans, and could induce accidents that humans could have avoided.…”

Section: Introductionmentioning

confidence: 99%

Robust Control for Over-Actuated Vehicles

Kissai

Monsuez

Martinez

et al. 2019

2019 IEEE Conference on Control Technology and Applications (CCTA)

View full text Add to dashboard Cite

Many car manufacturers are equipping their vehicles with advanced systems to win the race towards autonomous vehicles. In order to ensure multiple objectives, different embedded systems influencing differently the same physical variable are often implemented in the same vehicle. Most of car manufacturers tend to activate one system at a time, especially when the low-level control of actuators is developed by different equipment suppliers, making the inner dynamics uncertain. However, this limits the potential of the vehicle. This paper discusses the feasibility of coordinating these systems when activated at the same time, while ensuring an acceptable robustness regarding the uncertain dynamics. Results showed good performance in severe maneuvers when combining robust control synthesis and optimization-based control allocation algorithms. Therefore, several embedded systems may be activated at the same time to expand the overall potential of the vehicle and deal with more difficult situations.

show abstract

Efficient statistical validation of machine learning systems for autonomous driving

Cited by 13 publications

References 10 publications

Probabilistic Risk Assessment of an Obstacle Detection System for GoA 4 Freight Trains

Probabilistic Risk Assessment of an Obstacle Detection System for GoA 4 Freight Trains

Software Verification and Validation of Safe Autonomous Cars: A Systematic Literature Review

Robust Control for Over-Actuated Vehicles

Contact Info

Product

Resources

About