Welcome, Professor Niao He

Professor He, welcome to ETH Zurich! What are your current research interests?
My research lies in the interface of operations research and machine learning, with a primary focus on the algorithmic and theoretical foundations of data-driven decision-making problems. The goal of my research is to leverage data, optimisation and learning towards making provably efficient, reliable and intelligent decisions for next-generation AI. My principal research interests include large-scale optimisation, optimisation under uncertainty, reinforcement learning and probabilistic inference. I am also interested in developing machine learning models and algorithms for interdisciplinary applications in operations management, robotics, healthcare analytics, etc.

What is the impact of your research on society?
Machine learning (especially deep learning and deep reinforcement learning) has enjoyed phenomenal popularity and made breathtaking progress in practical applications in the past decade. Besides the unprecedented data volume and computing power, success in this field is also heavily reliant on efficient heuristics, hyperparameter tuning, empirical exploration and architecture engineering. As a result, even the state-of-the-arts often suffer from limited theoretical understanding, model interpretability and resiliency, thus hampering their deployment in the uncertain real world. My research focuses on developing foundational principles for machine learning from the fresh perspective of optimisation. This opens a doorway to principled, versatile and robust algorithms with both theoretical and practical efficiency. I believe it is imperative to bridge these gaps between theory and practice in order to help accelerate trustworthy AI and broaden its applicability in our everyday life.

Where were you working before you came to ETH Zurich?
Before joining ETH Zurich, I had been an Assistant Professor at the University of Illinois at Urbana-Champaign since 2016. Prior to that, I received my PhD from Georgia Tech in 2015.

Which courses will you be teaching at ETH?
I will be teaching undergraduate and graduate courses in data science and machine learning. I am also looking forward to developing new topic courses in big data optimisation and reinforcement learning, to bring students nearer to the frontiers of research in these fields.

What are your first impressions of Switzerland and ETH Zurich?
While I was not able to visit Switzerland and ETH Zurich beforehand due to the pandemic, my virtual visit to the country and the university has already been mind-blowing – the magnificent scenery, diversity and, most importantly, the enthusiastic and inclusive culture here. The faculty and staff at ETH have been incredibly supportive and helpful, making our transition from one continent to the other so easy. I am very excited to be part of this amazing group.

What advice would you give to students who are just starting out in computer science?
Computer science is a fast-moving field. The journey of PhD research in this area is a process of reinforcement learning by itself. Here are three pieces of advice borrowed from the wisdom of reinforcement learning algorithms:

1. Balancing exploration and exploitation: read broadly, but think deeply and critically;
2. Optimism in the face of uncertainty: aim for the long-term reward and big goal, and don't let small setbacks defeat you;
3. Never settle for suboptimal solutions: keep improving and always be making progressive steps.