Paper Accepted by ISSTA 2021

One paper entitled Attack as Defense: Characterizing Adversarial Examples using Robustness is accepted by ISSTA 2021. This work proposes to use ensembled attack strategies to defense adversarial examples (potentially from adaptive attacks) based ...

SIGSOFT Research Highlights

Pleased to share that our ICSE 2020 work on testing the fairness of deep learning models has been selected as ACM SIGSOFT Research Highlights of year 2020. Details can be found here.

Paper Accepted by TACAS 2021

One paper entitled Improving Neural Network Verification through Spurious Region Guided Refinement is accepted by TACAS 2021. This work proposes a spurious region guided refinement framework for over-approximation based neural network verificatio...

Paper Accepted by ICSE 2021

One paper entitled RobOT: Robustness-Oriented Testing for Deep Learning Systems is accepted by ICSE 2021. This work proposes a novel testing framework other than neuron coverage metrics to better improve the robustness of deep learning systems by...