Tuesday, May 25, 2010

Honoring Patrick Winston

Patrick Winston is a professor of artificial intelligence at MIT. Here's been here, in fact, since 1970, and during that time has honed his teaching to a science. An excellent speaker, he delivers his "How to Speak" talk in 6-120 annually during IAP, which is always oversubscribed and infamous for turning away fifty or so people at the door.


People fill the chairs and steps of 6-120, and even crowd around Winston's feet. (Jan 2009)

His website reveals him to be a man of many endeavors, from research to writing a community blog to befriending Will.i.am of the Black Eyed Peas.

I took 6.034 (Artificial Intelligence) with him last fall, and had such a good experience that when the MIT awards nomination email came around in February, I immediately thought of Winston. I was familiar with the Everett Moore Baker Memorial Award for Excellence in Undergraduate Teaching, since another of my favorite professors, Eric Hudson of Physics (who sadly did not get tenure and is leaving MIT - more about him in a later post), also won this award, in 2008. The Baker award is the only institute-wide teaching award based entirely on nominations from students.

Luckily, throughout 6.034, I had noticed a lot of details that made Winston's class great. When it came time to write the nomination letter, I just sat down and cranked out over five 1.5-spaced pages. Over the next two weeks, I would edit and revise it twice.

I then hunted down people to sign the letter. (It would have been acceptable with only my signature, but more names makes a stronger case.) "Hunt" is not too far from the truth. At the beginning of March, I sorted through the 6.034 class list and emailed all the people I knew, however tenuously, asking them to submit personal anecdotes and sign the letter. At one point, I posted myself outside the entrance to Winston's spring course, 6.803 (The Human Intelligence Enterprise), and intercepted people as they entered the class. I asked my 6.042 TA's to sign, and even the doctorate candidates at my new UROP (in the field of artificial intelligence, no less). I was.. an MIT student on a mission. In the course of my mission I found out that a nomination for Winston had been previously attempted, but had failed, which only made me feel more like this would be the year.

Okay, so I'm getting somewhat dramatic here, though admittedly this was a minor production. At the end, I delivered the nomination letter to the awards office (which is strangely nestled in the DuPont Athletic Center), and hoped the woman I gave it to would take good care of the packet. It occurred to me I hadn't made a copy of the signatures page.

I then kind of forgot about the whole business. I marked May 4, the date of the awards convocation, on my calendar, but I didn't have time to attend that day. Not until yesterday, after finals and everything, did I suddenly remember and check the awards website.


Winston won! My term is complete.



---------
Below is the nomination letter for Winston. Full names of students have been abbreviated.
It is long (you have been warned) but provides a good review of 6.034 for anyone looking to take it.

---------
March 10, 2010
Dear MIT Awards Committee,


I took Professor Patrick Henry Winston’s Artificial Intelligence course, 6.034, this past fall semester, and enjoyed it greatly. I’m not alone, as the list of signers below attests. Each of us has either experienced Professor Winston’s teaching through 6.034 or 6.803 – The Human Intelligence Enterprise, or both. Together, we enthusiastically nominate him for the 2010 Everett Moore Baker Award for Excellence in Undergraduate Teaching.


Professor Winston is a great teacher foremost because of his pedagogy, which he has honed to an art. He has crafted a potent combination of classroom management, lecturing prowess, and course design, all infused with a genuine personal touch. I will describe to the best of my ability the experience of being in his class.

Every morning before 10 AM on lecture days last fall, music could be heard streaming from room 10-250. Often it was the Rolling Stones, but one week it was traditional Japanese drumming by the Kodo Drummers; the next week, hip-hop by the Black Eyed Peas, juxtaposed with a picture on the projector screen of Professor Winston standing next to one of the band members. Besides waking students up and setting the mood (the playlist was related to the day’s lecture topic), the music made sure class always started on time, as it was programmed to fade out at exactly five past the hour. Professor Winston also told us that rock n’ roll pumped him up for lecture.

An outsider walking into 6.034 would have made a striking observation: not a single student ever had a laptop open. Professor Winston told us on the first day that he did not allow open laptops in lecture because they were a distraction to students and to him. So whether they liked it or not, students showing up to lecture were thus encouraged to pay attention.

Professor Winston engaged students not only as a group, but also individually, during lecture. Periodically, he would pitch a question to the sea of faces before him. Especially during the first few weeks, his questions were sometimes followed by the silence of his shy audience. But this was a minor obstacle—“Benjamin,” Professor Winston would say, addressing a student in the crowd by name, “what do you think?” The first time it happened, I was shocked. Professor Winston knew our names! There were at least two hundred students in the class, and yet he demonstrated throughout the semester that he knew who we were (at least those of us who showed up to lecture). I discovered this firsthand: as I passed by him while heading out of the second lecture, he squinted at me and said, “You’re Jessica, right?” After that, whenever I saw him in at the Stata Center, I always said “Hi,” and he always greeted me by name.


So far I’ve talked about the environment surrounding Professor Winston’s lectures. But the lectures themselves deserve praise. From the beginning of term to the end, each lecture was masterfully well-organized and easy to follow.

At the end of each Independent Activities Period at MIT, Professor Winston gives a well-known talk entitled “How to Speak,” which attracts a hundred more people than can be comfortably seated in the hall in which it is delivered, 6-120. People end up squeezing into the aisles, sitting on the steps and even spreading out onto the floor at the front of the room, under Professor Winston’s nose. And every year, people are turned away at the door because there is just not enough room.

The fact that Professor Winston can give such a popular talk about “How to Speak” indicates his own ability to deliver a great talk. And sure enough, he follows his tenets of speaking during his lectures. He always begins his lectures by promising us that by the end of the hour, we will have learned A, B, and C, and be able to do X, Y, and Z. It’s a great hook.

His lectures flow and are easy to follow. In fact, I’ve heard some students criticize his 6.034 lectures for being too simplistic. But I think this is the wrong way to look at it. Artificial intelligence is by no means an easy subject—researchers have spent decades on the matter. To make the subject seem “simple,” then, means that Professor Winston can break complicated material down into a digestible form, giving us the big ideas. I have certainly been in classes at MIT where I was totally confused after lectures, only to realize afterward that, “Hey, this concept isn’t actually that complicated. So why did the professor explain it in that convoluted way?”

In contrast, in 6.034, Professor Winston found very clear ways and good examples to illustrate his points. He sometimes brought in props, such as cardboard blocks for a lecture on machine vision, to help us visualize problems. When he drew pictures on the board, his technique was excellent—his 3-D blocks came out in wonderful perspective. At the end of each lecture, he almost always provided a demonstration of a computer program that implemented the lecture topic, so that we could see the theory in action. I still remember the major concepts from 6.034 well now, because I can remember the examples he gave.


Professor Winston’s pedagogical excellence extends beyond his lectures. The class structure and grading policies also are key elements to the success of 6.034. There are four different styles of meetings in the course: there are the usual lectures and recitations, but also tutorials and “mega-recitations.” Tutorials are smaller sections of about eight to ten students led by a TA, in which we solidify our understanding of the topics, feel free to ask any question in this more private setting, and solve problems. Mega-recitations are lecture-sized recitations in which more problems from previous years’ quizzes are worked out. Each of these class settings is useful for different reasons, and together with the programming assignments they give every student plenty of opportunities to clear up any confusion and to see the material in many ways, which are keys to comprehension.

Professor Winston has also devised an innovative grading policy that emphasizes learning and reduces student anxiety. Instead of having one or two stressful high-stakes exams, the material is divided into five smaller quizzes. There are four in-class quizzes throughout the term, and the fifth quiz appears on the final exam. However, sections corresponding to the first four quizzes are found on the final exam too, and they give students a second chance at testing the material. The highest score that students get on each section—either the first time on the in-class quiz, or on the final—becomes their score for that section.

This grading policy makes the class more enjoyable, since students don’t see one poor quiz during the term as a big setback—they know they can redeem themselves on the final. Furthermore, the grading policy takes the pressure off individuals to have to learn at the same pace as everyone else. If a student needs a whole semester for the material to sink in and make sense, but can demonstrate understanding on the final, then that student can also succeed in the class. Students who do perform well on the quizzes the first time around don’t complain either—they can ignore the parts of the final that they don’t need to take, and concentrate their efforts on Quiz Five and any parts that they do need to make up for.

Another aspect of the grading policy is that quizzes are graded on a scale from 1 to 5. More accurately, they are graded out of 100 points first, and then assigned a score of 1 to 5 based on cutoffs. The cutoffs, which are unique for each quiz, are determined by the staff, who decide what scores indicate “thorough understanding,” “adequate understanding,” and so forth, given allowances for reasonable mistakes. Class averages are not calculated—they aren’t needed, because the number awarded to the student indicates how well the student has demonstrated understanding. This grading scheme de-emphasizes comparison between students and focuses students on their own success. Also, it is made clear at the beginning of the semester what average scores are needed to earn an A, B, C, etc. in the course. So students know exactly how well they are doing in 6.034 at any given time. Another effect of this grading scheme is that it clearly does not grade students on a curve. Anyone who obtains the average score for an A will get an A. The grading policy is fair, clear, and emphasizes learning while decreasing stress and competitiveness.


Professor Winston not only covers the curriculum, but goes beyond to make an impact on our lives. He not only instructs, but is honestly concerned about students’ well-being. Students can sense that.

For example: In every lecture, Professor Winston imparted to us “Big Ideas,” which he often named. One of the first ideas was the “Rumpelstiltskin Principle,” which said, “If you can name an idea, you have power over it.” But one of these Big Ideas was also: “Protect your sleep like your wallet.” Winston devoted a good fifteen minutes of lecture one day to talk about the importance of sleep to learning. It was relevant to the class because we students were learning about machine learning… or at least attempting to. (It also gave us insight into human learning, which is boosted by adequate sleep.) Professor Winston knows the reputation of MIT students of working hard and sleeping little. So he showed us charts of people’s mental performance correlated with the numbers of hours of sleep they got, and I saw that I was losing about thirty percent of my peak performance based on the amount of sleep I was getting. I knew I didn’t feel great every day, but that lecture actually scared me a little. I’ve been trying hard to sleep seven hours a night on average since then, and in general I feel a lot better.

Another example: in the lecture before Thanksgiving, for which he expected lower attendance, Professor Winston continued his annual tradition of giving a talk on “the Artificial Intelligence Business,” or, as he quipped, “how to get famous.” He told us about running his own company and explained the business lessons he learned. This lecture demonstrated that Professor Winston cared not only about the technical knowledge in our heads, but also about our success in the future.

Personally, 6.034 inspired me to get a UROP in artificial intelligence, and I’ve been happily working on it since. Others have been similarly inspired; a girl who was in the same 6.034 recitation as me now sits beside me at my UROP.


I’ve described Professor Winston’s teaching. Aside from being a great teacher, though, Professor Winston is a genuinely thoughtful and understanding person. You can tell from the little things he does. He regularly photographs his blackboard notes after his lectures for students who are absent for good reasons. And when a friend of mine was injured and needed an extension on an assignment, he approached Professor Winston to ask for one, and Professor Winston replied, “Of course! But first, are you okay?”

Professor Winston is an active contributing author on the MIT Alumni Association’s Slice of MIT blog (http://sliceofmit.wordpress.com/author/patrickwinston/). There, you’ll find his lighthearted, thoughtful accounts of his life at the Institute, and you can see firsthand his close association with students, as he writes about attending an MIT Dramashop production of the play Rossum’s Universal Robots, and driving to Endicott College to cheer on a former 6.034 student and star MIT football player at the last game of the season. His tales reveal his perennial knack for humor and sparkle with his hallmark spunk and wisdom.

As I mentioned before, I am not the only supporter of Professor Winston’s teaching. Here are some comments from other students in his 6.034 class this past fall:

“Every day before 6.034, Winston would take an hour preparing his lecture on the chalkboards in the Stata Center. He would prepare everything on the chalkboards down to the slightest detail, and it was evident his preparation worked: I understood all the material. One day, I missed class and went to Stata to write down Winston's chalked notes. Winston happened to walk by me, smiled, and said that the notes were exactly the same as those he presented earlier during lecture. Now that’s consistency!”
—R. M.

“I remember the first day of Prof. Winston’s class. I had heard so many good things about him; it was like going to a well-advertised movie. Unlike well-advertised movies, however, Winston lived up to his hype. From the very first greeting to the last farewells, Winston had a great positive effect on my studies.”
—A. S.

Whenever I talk to fellow undergraduates about 6.034, the introductory Artificial Intelligence class, I get a whole bunch of enthusiastic responses and a few people who didn't like the course. When I ask those who didn't enjoy it whether they went to lecture, the answer is always no. That explains everything - 6.034 is a course made popular not by its material, but by its lecturing professor, Patrick Winston. Professor Winston is the most inspiring MIT professor I've ever had the pleasure of having; during his classes, he speaks not only about course material, but imparts upon us wisdom and life lessons. In 6.034, I did not learn only about Artificial Intelligence - I learned about how to improve my focus, what lack of sleep does to my brain functioning, how to give effective presentations, how to approach a problem when its current representation has limitations, how to found an effective startup...the list goes on and on. It is clear from simply listening to Professor Winston talk that he cares more about his students than just what they learn about Artificial Intelligence - he believes in us, believes that we have the capability to do great things, and goes out of his way to inspire us and give us the requisite tools to be successful. And he is wildly successful. After taking 6.034, I now sleep more and take better care of myself. I find myself referencing my 6.034 notes whenever I have to speak in front of groups of people. And I talk to myself more, because talking to oneself helps one focus, and I find that my performance has improved with ease. It's all of these little things I took away from 6.034 that makes Professor Winston so incredible - sure, I know how Dijkstra's algorithm works now, but any professor can do that. What's important is that Professor Winston has, through a single class, managed to impact and improve my daily life beyond the scope and timeframe of that course. And he has inspired me to seize opportunities and do something great - something I'd never believed until he believed that I could - and I can.
—K. Q.

Professor Winston does an excellent job of passing on his enthusiasm for the subject to his students. In addition to teaching us about various AI concepts and the stories behind them, he also gave some very memorable lectures on the importance of sleep, what life is like when you're trying to run an AI startup, and current research in sensorimotor and language learning that inspired me to pursue further classes and a UROP in AI.
—E. C.

“Professor Winston has become one of my role models here at MIT and an inspiration for me to learn. The knowledge that he passes to us not only helps us learn about artificial intelligence, but also about how to be successful in any of our endeavors. I cannot think of a better professor to be the recipient of the Everett Baker Award.”
—A. M.

I can think of no other professor who has his students' best interests and this institute's best interests at heart more than Professor Winston. When I was a sophomore, I had issues in my personal life that were affecting my academic work. As my advisor, Professor Winston gave me advice on how to move forward and made sure I was doing well with several follow-up communications. He is one of the reasons I am where I am today. I'm now taking The Human Intelligence Enterprise with Professor Winston; his insight and experience make it one of the most interesting and useful classes I've ever taken.
—E. F.


Professor Winston’s impact is not limited to this term—some of the signers below are alumni who took 6.034 in previous years. In fact, Professor Winston has been teaching at MIT since 1970. We can only wonder how many students he has inspired in the past forty years. We believe that Professor Winston is a perfect candidate for the Everett Moore Baker Award for Excellence in Undergraduate Teaching, and we would be gratified to see him honored in 2010.

Sincerely,
.....

No comments:

Post a Comment