Undergraduate Research Assistant
CurrentPreparing a paper: Ben Mckenzie, Xiaohan Chen, Wotao Yin and Zhangyang Wang, "Safeguarded Training for Learned Optimization”, targeting at ICLR 2021. Working closely with Dr. Atlas Wang on the research of Learning-based Optimization. The focus is on safe-guarded training, a new algorithm that we proposed for addressing the out-of-domain adaptation of learned optimizers. Duties include:- Using self-learned real analysis, optimization, and advanced machine learning skills.- Implementing and training various in NN-based sparse coding models (LISTA, etc.) with PyTorch and Sklearn- Using Numpy and Pandas to create and manage experimental data.