This article is a translation of a Chinese report by the WLA Forum, with excerpts.
You might find it difficult to understand machine learning and its related terms such as probability theory, statistics, and Bayesian networks, but you probably use E-mail, search engines, face recognition, and e-commerce platforms on a frequent basis. Also, self-driving and other versatile prevalence of artificial intelligence (AI) never fail to amaze. The underlying logic of these applications derives from theories laid down by Michael I. Jordan over the past 40 years.
Jordan is a professor in the Department of Electrical Engineering and Computer Science and the Department of Statistics at the University of California, Berkeley, and is the first AI scientist who was elected as a member of the National Academy of Sciences, the National Academy of Engineering, and the American Academy of Arts and Sciences. He is also one of the pioneers in machine learning, and considered "the root directory" in the field of AI.
On September 29, Jordan became the first recipient of the WLA Prize in Computer Science or Mathematics.
John Hennessy, Chairman of the WLA Prize Selection Committee in Computer Science or Mathematics and 2017 Turing Award Laureate, stated during his speech to honor Jordan, "Professor Michael Jordan has built connections between statistics, graphic modeling, probabilistic methods, and machine learning. These connections not only are conducive to the development of machine learning but also expand and improve relevant research work."
In recognition of Jordan's contributions to machine learning, he singlehandedly won the prize.
From Research Psychologist to Machine Learning Forerunner
Jordan was born a baby boomer in the late 1960s and spent his early years in Louisiana, USA, a state with unique local customs converging American, French, and African cultures. Growing up, he was full of curiosity and showed great interest in the study of philosophy, cultural beliefs, and ways of thinking, reports said.
During his undergraduate study, he was inspired by Bertrand Russell's autobiography, who explored thought is a logical and mathematical process, according to a report released by IEEE Spectrum, a flagship publication of the IEEE, in 2021.
"Thinking about thought as a logical process and realizing that computers had arisen from software and hardware implementations of logic, I saw a parallel to the mind and the brain," said Jordan in the IEEE Spectrum report. "It felt like philosophy could transition from vague discussions about the mind and brain to something more concrete, algorithmic, and logical. That attracted me."
As he recalled, despite being intrigued by machine learning, he already felt at that time that the deeper principles needed to understand learning were to be found in statistics, information theory and control theory. This may offer a clue or two on how he developed his original thinking.
Jordan left the Massachusetts Institute of Technology (MIT), and joined the University of California, Berkeley in 1997 at the invitation of Peter Bickel, a statistician there, to pursue his studies in probability and statistics, according to his profile in 2013 by the Proceedings of the National Academy of Sciences (PNAS), a peer reviewed journal of the National Academy of Sciences.
Jordan explained the reason of leaving the MIT was mainly because there was no statistics related majors in the MIT at that time. He holds a strong belief that despite the exciting development of computer science today, probability and statistics require further study as there remains a disconnection between computation and interference.
During his research career, Jordan has identified the connections between machine learning and statistics, and raised awareness of the importance of Bayesian networks in the field of machine learning. He has also built connections between statistics, graphical modeling, probabilistic methods, and machine learning, and particularly laid a foundation in computer science and mathematics for machine learning.
"Developing a safe and reliable theoretical framework for technical applications is a major challenge in machine learning. Only with this framework can we truly understand how these technologies work and whether the results are reliable. Jordan’s research has laid the underpinnings and structure of machine learning and made the cornerstone of artificial intelligence, " said Hennessy.
Jordan is reputed as an all-rounder scientist in media reports. He has also branched out of the academia, aiming to build systems in forecasting, distribution, economy, transportation, commerce, law, and entertainment as solutions to practical problems involving optimization, statistical interference, algorithms, and platforms.
Building the Human Resource Underpinning of AI
In Hennessy’s words, Jordan has made further contributions to machine learning and AI by developing truckloads of outstanding and creative human resources in this field.
In addition to authoring numerous textbooks, Jordan has instructed more than 80 doctoral students and 60 postdoctoral researchers, and all of them have become either professor in leading academic institutions worldwide or influential industry leaders.
There are many well-known figures among them: Yoshua Bengio, 2018 Turing Award Laureate and a professor at the University of Montreal; Zoubin Ghahramani, an authority in Bayesian learning and a professor at Cambridge University; David M. Blei, an authority in LDA and a professor at Columbia University, and Andrew Ng, a professor at Stanford University.
Following his philosophy, his disciples have been committed to seeking connections between relevant fields. Among them, Andrew Ng, a highly reputable AI expert in China, was not only the Director of the Stanford AI Lab, a co-founder and head of Google Brain, also the former Chief Scientist at Baidu.
Jordan enjoys talking with scholars and students from all over the world to discuss research trends, researchers’ mentality, student training among other topics, and has visited China several times. He spent several years as a distinguished visiting professor at Tsinghua University, and as of now, an honorary professor at both Tsinghua University and Peking University.
There are myriads of interesting stories about Jordan on Zhihu, a question-and-answer website for Chinese researchers. For example, he has long left his students to their own devices. One of his PhD students wrote in the dissertation acknowledgment, "He gave up on me at the beginning of my doctoral journey, so I had to develop my independent thinking and research skills." And a multi-instrumentalist as he is, Jordan is the lead drummer in the professors' band of the Department of Electrical Engineering and Computer Sciences at UC Berkeley.
In 2023, he will also teach the course "Theoretical Foundations of Learning, Decisions, and Games" at UC Berkeley.
"The Real AI Revolution Hasn’t Happened Yet"
In the face of the global AI boom, Jordan as a long-time forerunner in this field remains calm. Artificial-intelligence systems are nowhere near advanced enough to replace humans in many tasks involving reasoning, real-world knowledge, and social interaction, noted Jordan in the IEEE Spectrum report. Though human-like competence are spotted in performing rudimentary pattern recognition tasks, but at the cognitive level they are merely imitating human intelligence, not engaging deeply and creatively, said Jordan.
In 2018, Jordan published an influential article Artificial Intelligence—The Revolution Hasn't Happened Yet on AI developments in Harvard Data Science Review. His comment on the current AI hype argues that the majority of what is labeled AI is actually machine learning, and that the real AI revolution hasn’t happened yet.
Jordan reckons that the so-called intelligence at present is a data algorithm aggregated from parameters, and it can only replicate, imitate, and simulate human behaviors. It is not actual intelligence.
Bringing computers and humans together in ways that enhance human life is a major challenge on our hands. That's why Jordan believes that interdisciplinary collaboration is very important. He stressed in the article that a full scope understanding on machine learning is more about the need for humans to understand and shape technology as it becomes ever more present and influential in their daily lives. Moreover, during the process of understanding and shaping the technology, there is a need for a diverse set of voices from all walks of life, not merely a dialog among the technologically attuned. Focusing narrowly on human-imitative AI prevents an appropriately wide range of voices from being heard, especially perspectives from the social sciences and humanities.
He believes that human well-being should not be something that can wait after the development of technology. "In the current area, we have a real opportunity to conceive of something historically new: a human-centric engineering discipline," he writes.
Jordan has confirmed his attendance at the Award Ceremony of the inaugural World Laureates Association Prize (WLA Prize) at the opening ceremony of the 5th World Laureates Forum (WLF) this November, followed by a host of featured forum sessions.
