Causality-based neural network repair
Authors:
Open Access
- 21 May 2022
- conference paper
- Published by Association for Computing Machinery (ACM)
- p. 338-349
Abstract
Neural networks have had discernible achievements in a wide range of applications. The wide-spread adoption also raises the concern of their dependability and reliability. Similar to traditional decision-making programs, neural networks can have defects that need to be repaired. The defects may cause unsafe behaviors, raise security concerns or unjust societal impacts. In this work, we address the problem of repairing a neural network for desirable properties such as fairness and the absence of backdoor. The goal is to construct a neural network that satisfies the property by (minimally) adjusting the given neural network's parameters (i.e., weights). Specifically, we propose CARE (CAusality-based REpair), a causality-based neural network repair technique that 1) performs causality-based fault localization to identify the 'guilty' neurons and 2) optimizes the parameters of the identified neurons to reduce the misbehavior. We have empirically evaluated CARE on various tasks such as backdoor removal, neural network repair for fairness and safety properties. Our experiment results show that CARE is able to repair all neural networks efficiently and effectively. For fairness repair tasks, CARE successfully improves fairness by 61.91% on average. For backdoor removal tasks, CARE reduces the attack success rate from over 98% to less than 1%. For safety property repair tasks, CARE reduces the property violation rate to less than 1%. Results also show that thanks to the causality-based fault localization, CARE's repair focuses on the misbehavior and preserves the accuracy of the neural networks.Keywords
All Related Versions
Funding Information
- Huawei International (TC20210714014)
This publication has 33 references indexed in Scilit:
- Repairing Decision-Making Programs Under Uncertainty2017
- FairTest: Discovering Unwarranted Associations in Data-Driven Applications2017
- Using deep learning to investigate the neuroimaging correlates of psychiatric and neurological disorders: Methods and applicationsNeuroscience & Biobehavioral Reviews, 2017
- Estimating Conditional Average Treatment EffectsJournal of Business & Economic Statistics, 2015
- Certifying and Removing Disparate Impact2015
- FaceNet: A unified embedding for face recognition and clustering2015
- Man vs. computer: Benchmarking machine learning algorithms for traffic sign recognitionNeural Networks, 2012
- What causes a system to satisfy a specification?ACM Transactions on Computational Logic, 2008
- Particle swarm optimizationSwarm Intelligence, 2007
- Parameter selection in particle swarm optimization1998