Sanmi Koyejo, an assistant professor of computer science at Stanford University, is a leading figure in the field of trustworthy AI research. His work focuses on the intersection of machine learning, scientific discovery, and the complex question of how to trust AI systems. Koyejo's journey into AI began unexpectedly, as he initially pursued a career in electrical engineering, building electronics and working on control systems. However, during his graduate studies, he discovered the exciting world of machine learning and quickly shifted his focus.
Koyejo's research group is dedicated to three main areas: understanding AI systems, building trustworthy AI, and applying AI to real-world problems, particularly in science and healthcare. One of the reasons Koyejo is drawn to astronomy is the unique nature of the problems it presents. Unlike other fields where AI excels, astronomy operates within a single universe, making it challenging to create new datasets for testing. Instead, astronomers rely on combining observations with physical understanding and scientific intuition, which Koyejo finds exciting.
A key aspect of Koyejo's work is the distinction between doing well on a test and doing science. While benchmarks are useful for evaluating AI systems, he argues that passing these tests does not necessarily mean the AI can solve larger problems. His research aims to bridge the gap between impressive demonstrations and reliable real-world performance, emphasizing the importance of evidence and understanding.
Koyejo's plenary talk highlights the need for scientists to actively shape AI tools. As AI becomes more integrated into scientific work, researchers should not be passive users but should contribute to the development of these tools. They understand what constitutes evidence, which mistakes matter, and what makes a result trustworthy, and these insights are crucial in shaping AI's role in science.
For students, Koyejo offers valuable advice. With technology enabling rapid production, it's essential to focus on understanding rather than just generating outputs. Mentorship plays a vital role in developing judgment, perspective, and taste, which are increasingly valuable skills in a rapidly changing landscape. Koyejo encourages students to explore widely and consider the impact they want to make, whether in fields like astronomy or other areas.
In conclusion, Sanmi Koyejo's work at the intersection of AI and science is both exciting and thought-provoking. His research aims to improve the reliability and trustworthiness of AI systems while applying them to real-world problems. As AI continues to evolve, Koyejo's contributions will undoubtedly shape the future of this rapidly developing field, and his insights will be invaluable for scientists and students alike.
