Horace P. Yuen (1946-2025)

A farewell to my PhD advisor.

Three weeks ago I received the sad and shocking news that my PhD advisor, Horace P. Yuen, passed away at the age of 79. Everyone always says on such occasions how much they had been influenced by their advisor. I won’t be an exception to this. Horace’s influence on me has indeed been truly profound and extended from the usual academic matters to philosophical reflection and a general attitude toward life and human affairs.

My first encounter with Horace was in 1998, before I even started entertaining any thoughts of pursuing a graduate degree. I was an Electrical Engineering junior at Northwestern University. At that time I was mainly interested in optoelectronics and fiber optics and was eyeing an engineering career in industry, at a place like Lucent or Corning. There was the usual chatter among our cohort about which classes to take and with whom, and, when it came to the second required course in electromagnetics, the overall sentiment was something like, “don’t take it with Yuen, he’s an old school guy with a PhD from MIT and he expects you to know all this math.” Some part of me must have reacted like, “oh yeah? I think I could totally do it,” so I registered for his class. That was the turning point.

He would stroll into the lecture room, carrying a manila folder with two or three handwritten sheets of notes. He barely looked at them when he lectured. He wrote on the board in perfect longhand and would often add personal anecdotes about science and scientists or philosophical digressions about the difference between physics and engineering. One of his frequent sayings was, “mathematicians care about rigor, engineers care about precision.” What others had been saying about his demands on our mathematical knowledge was not entirely accurate—while he certainly emphasized mathematical sophistication, to him it was not about having lots of tricks at your disposal, but more about your ability to drill down to the conceptual core of things, even when thinking about something as basic as the definition of angle. Compared to your typical class where the emphasis was about grinding out problem sets, this was a refreshing experience.

My interest in photonics led me to quantum electronics and quantum optics, and Horace was the one to talk to about these things. He was legendary for his 1976 work on squeezed states of light, and also because he drove an opalescent maroon Jaguar with the license plate that said “CHRONON” and was seriously into competitive ballroom dancing. In my senior year, I signed up for his graduate course on information theory and was a regular in his office, just coming in whenever I had some half-baked thought or a question about some reading I was doing at the time. His door was always open, and he would never turn anyone away. That was a double-edged sword though, because he was also very direct and did not mince words. At some point, when I came in with another one of my half-baked ideas, he looked at me and said, “Max, I don’t think you are capable of abstract thought.” Did that discourage me? Not at all. He must have sensed that because the next week he handed me a copy of Michael Artin’s Algebra and said, “I got this for you. I learned abstract algebra from Artin when I was an undergrad at MIT, this is the way to learn it.”

Horace’s information theory course was challenging but incredibly rewarding not just because of the original way he presented the material (his introduction of Shannon entropy as “asymptotic combinatorial complexity” will forever be etched in my memory), but also because of all the stories he would tell. He was first a student and then a researcher at MIT when Shannon was on the faculty there, and his insights into the history and the conceptual development of information theory were remarkably deep. He gave us a lot of additional reading, including Shannon’s original papers, Jim Massey’s “Information theory, the Copernican system of communication,” and David Slepian’s “On bandwidth.” Most importantly, he would continually emphasize the importance of not losing sight of practical engineering considerations when working with mathematical models.

When I eventually became his PhD student, I had complete freedom to work on whatever I wanted. Horace was always there to suggest relevant references. He had an incredible memory and a quick way of seeing straight to the heart of the matter. At one point, when I became interested in reliable computation with unreliable circuit elements after having read von Neumann’s famous work on this problem, I asked Horace whether it made any sense to talk about something like “computation capacity” in analogy to Shannon’s channel capacity. He answered right away, “sure, take a look at Winograd and Cowan’s Reliable Computation in the Presence of Noise.” He had zero tolerance for hype and self-promotion. Every time I would come into his office and say something like, “have you seen this recent paper by so-and-so? they did such-and-such,” he would often reply, “well, they claim that they did it, is it really true?” This skeptical attitude is one of the things I learned from him. And, of course, he remained true to his nature. When one of his frequent collaborators was visiting from Italy, the three of us were in Horace’s office, and, in the middle of a heated discussion about some derivation on the board, Horace exclaimed to him: “Come on, how is it that you don’t understand this? Even Max understands it!” And yet, despite all this, he has been unwavering in his support and encouragement.

We remained in touch after I defended my PhD in 2002, and even later when I switched fields from quantum information to machine learning and control. He would take me out to dinner either at one of his favorite Cantonese restaurants in the Chicago area or at one of the upscale establishments in Evanston, and we’d have long conversations about whatever was on our minds. He was always open to learning new things, and he brought his skeptical attitude, his incredibly broad knowledge of a vast array of subjects, and his remarkable memory to everything. I don’t remember exactly how the subject of philosophy came up, but at one point I was telling him about my reading of MacIntyre’s After Virtue. Little did I know that, among many other things, he was deeply read in both continental and analytic philosophy, going back to the time when he and his ex-wife Sunny Auyang would attend philosophy seminars at MIT and study Heidegger and Husserl together. In that discussion we had about MacIntyre, he brought up Nozick’s Anarchy, State, and Utopia and Rawls’ Theory of Justice and had interesting critical takes on both and on their relation to MacIntyre’s views.

He retired in 2021, shortly after the COVID pandemic, and we kept having long discussions over email. He wanted to understand recent developments in machine learning and AI and how they relate to ideas in philosophy of mind. As usual, he was incredibly quick in the way he understood things and got straight to the core; his replies to me were full of references to Rudolf Carnap, Daniel Dennett, and Jaegwon Kim. He was unconvinced by Kim’s physicalism, but found his books to be the best defense of the physicalist view. Just as in the old days, he believed that electrical engineers, and not computer scientists or physicists, were best equipped to make sense of AI; he was unimpressed by Geoff Hinton’s alarmism but found Yann LeCun’s views congenial to his own perspective. In many of our exchanges, he stressed the importance of approaching things from what he called his “Buddhist/Daoist/scientific perspective.” He argued that a comprehensive view of the world and of our place in it requires all three: Buddhism in order to understand our inner mental experience as directly given, Daoism in order to grasp how our inner experience affects and is affected by the web of causal relations in the world, and science in order to maintain coherence between the Buddhist/Daoist manifest image and the modern scientific image.

In some our last exchanges, we were getting into philosophical issues of Buddhism and Daoism, and I wanted to get Horace’s perspective as I was making my way through Chad Hansen’s book. We went deeply into the “Happiness of fish” passage from The Zhuangzi, and my amateur interest was obviously no match for Horace’s decades-long study of both classical and modern Chinese philosophy. He had many, many things to say, with references to Schopenhauer (whom he considered one of the deepest interpreters of Buddhism) and to the philosophy of Xiong Shili, of whom I had never heard before. I wish I did not delay responding to his latest email, which turned out to be his last to me. Here is what he wrote in one of our exchanges:

The 'realism' description of humans, together with the added human complexities on top, seems primarily a result of evolution. This is a scientific worldview that I have to reconcile with Buddhism and Daoism and whatever basic thoughts I may have. Maestripieri quotes Darwin, "He who understands baboon would do more towards metaphysics than Locke". The challenge to my perspective as a whole is clear.

It is empty to claim, as Buddhists will in response, that human reality is 'not real'. Why would they even bother to respond then? They are concerned with something, including the 'unreal' realities, with all their life instructions and prohibitions. Although I can argue further for them, they are not authentic nihilists. Daoists, especially Zhuangzi, would dismiss all these from a transcendental disposition which is seemingly not nihilistic, but logically appears to be just an 'active' flip side of nirvana.

Farewell, Horace. I hope you are now happy and carefree, like those minnows in the River Hao.