Publication

The template is mobile first with a responsive design to ensure that your site looks stunning on every device. ## Get Started - 👉 [**Create a new site**](https://wowchemy.

We are pleased to announce that our paper on efficiency backdoor attacks and transferability attacks for human face recognition systems has been accepted for presentation at the prestigious Conference on Computer Vision and Pattern Recognition (CVPR) in 2023.

Our paper, DyCL, is accepted to ISSTA 2023. DyCL is a automatic tool that rewrites the dynamic neural networks for supporting existing DL compilers.

Our paper, DeepPerform, is accepted to ASE 2022. DeepPerform is designed to test the performance robustness of adaptive neural networks.

Our paper, NMTSloth, is accepted to ESEC/FSE 2022. NMTSloth is designed to test the performance degradation bugs in neural machine translation systems. We show NMTSloth can generate generate inputs that increase neural machine translation systems computational consumption upto 3153%.

Our paper, NNReverse, is accepted to IJCAI 2022. NNReverse is an automatic approach for reversing DNNs from AI programs.

The template is mobile first with a responsive design to ensure that your site looks stunning on every device. ## Get Started - 👉 [**Create a new site**](https://wowchemy.

The template is mobile first with a responsive design to ensure that your site looks stunning on every device. ## Get Started - 👉 [**Create a new site**](https://wowchemy.

The template is mobile first with a responsive design to ensure that your site looks stunning on every device. ## Get Started - 👉 [**Create a new site**](https://wowchemy.

Our paper, NICGSlowDown, is accepted to CVPR 2022. To the best of our knowledge, NICGSlowDown is the first work to evaluate the efficiency robustness of neural caption generation models.