Cho-Jui Hsieh, an assistant professor of computer science at the UCLA Samueli School of Engineering, has received a 2021 Okawa Foundation Research Grant for his work on machine learning for optimization. Hsieh is one of only seven recipients from the United States to be awarded the $10,000 grant for 2021.
Established in 1996, primarily by the late information technology pioneer Isao Okawa, the Okawa Foundation for Information and Telecommunications in Japan promotes and develops the field of information communications technology by providing awards and grants to researchers, and hosting relevant symposiums and workshops.
Hsieh’s research has the potential to impact the field broadly by solving a problem many large networks face — namely, how to efficiently tune optimizers.
For a typical network to run smoothly when handling large volumes of data, the system has to first address optimization problems. However, current optimizers require a lot of manual tuning and are prone to human errors. Hsieh’s research has centered on developing artificial intelligence that can more efficiently adjust the optimizers for each given task and potentially in new ways that humans have yet to develop.
In addition to the research supported by the Okawa Foundation, Hsieh has developed algorithms that reduced the machine-training time by an order of magnitude on distributed systems. For example, companies such as Google and Nvidia use one of his algorithms to train BERT (Bidirectional Encoder Representations from Transformers), a widely used gigantic natural language processing model, in under a minute.
Hsieh served as an assistant professor at UC Davis for three years before joining the faculty at UCLA in 2018. His many honors include a Google Research Scholar award, Samsung’s AI Researcher of the Year and a CAREER Award from the National Science Foundation.
Natalie Weber contributed to this story.