Computer Science students Ruochen Wang, Minhao Cheng, Xiangning Chen, and faculty Cho-Jui Hsieh, in collaboration with Xiaocheng Tang at DiDi Research, have received the Outstanding Paper Award from the International Conference on Learning Representation (ICLR 2021) for their paper “Rethinking Architecture Selection in Differentiable NAS.” Their paper identifies an underlying reason why current differentiable Neural Architecture Search (NAS) algorithms fail and provides a simple fix that consistently improves many existing differentiable NAS algorithms. The International Conference on Learning Representation (ICLR) is the premier conference for deep learning. Among the 860 papers accepted this year, the committee selected only eight papers for the Outstanding Paper Award. Follow the link for the full award announcement: