Yao Qin is a Research Scientist at Google Research. She received a PhD degree in Computer Science and Engineering at UC San Diego under the supervision of Prof. Garrison Cottrell in 2020. Her research focuses on improving robustness of machine learning. Due to her contributions to the robustness of ML models (10 first-author publications and over 1500 citations), She has been selected as EECS Rising Star at MIT, 2021. Yao interned at Brain Toronto Team advised by Geoffrey Hinton in 2019, and Google Brain, advised by Ian Goodfellow in 2018.
There are many robustness issues arising in a variety of forms while deploying ML systems in the real world. For example, neural networks suffer from distributional shift — a model is tested on a data distribution different from what it was trained on. In addition, neural networks are vulnerable to adversarial examples – small perturbations to the input can successfully fool classifiers into making incorrect predictions. In this talk, I will introduce how to improve robustness of machine learning models by building connections between different perspectives of robustness issues and bridging gaps between a wide range of modalities. As a result, seemingly different robustness issues can be tackled by closely-related approaches, and robust ML on multiple modalities and backbone architectures can converge to a common ground.
Please contact MPI-SWS Office Team for link information