Few-Shot Guided Mix for DNN Repairing

Xuhong Ren, Bing Yu, Hua Qi, Felix Juefei-Xu, Zhuo Li, Wanli Xue, Lei Ma, Jianjun Zhao

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

Although deep neural networks (DNNs) achieve rather high performance in many cutting-edge applications (e.g., autonomous driving, medical diagnose), their trustworthiness on real-world scenarios still posts concerns, where some specific failure examples are often encountered during the real-world operational environment. With the limited failure examples collected during the practical operation, how to effectively leverage such failure cases to repair and enhance DNN so as to generalize to more potentially suspicious samples is challenging, but of great importance. In this paper, we formulate the failure-data-driven DNN repairing as a data augmentation problem, and design a novel augmentation-based repairing method, which to the best extent leverages limited failure cases. To realize the DNN repairing effects that generalize to specific failure examples, we originally propose few-shot guided mix (FSGMix) that augments training data with the guidance of failure examples. As a result, our method is able to achieve high generalization to the collected failure examples and other similar suspicious data. The preliminary evaluation on CIFAR-10 dataset demonstrates the potential of our proposed technique, which automatically learns to resolve the potential failure patterns in the DNN operational environment.

Original languageEnglish
Title of host publicationProceedings - 2020 IEEE International Conference on Software Maintenance and Evolution, ICSME 2020
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages717-721
Number of pages5
ISBN (Electronic)9781728156194
DOIs
Publication statusPublished - Sep 2020
Event36th IEEE International Conference on Software Maintenance and Evolution, ICSME 2020 - Virtual, Adelaide, Australia
Duration: Sep 27 2020Oct 3 2020

Publication series

NameProceedings - 2020 IEEE International Conference on Software Maintenance and Evolution, ICSME 2020

Conference

Conference36th IEEE International Conference on Software Maintenance and Evolution, ICSME 2020
CountryAustralia
CityVirtual, Adelaide
Period9/27/2010/3/20

All Science Journal Classification (ASJC) codes

  • Software
  • Safety, Risk, Reliability and Quality
  • Modelling and Simulation

Fingerprint Dive into the research topics of 'Few-Shot Guided Mix for DNN Repairing'. Together they form a unique fingerprint.

Cite this