Push Singh

We had had some trouble with the wiki database, but this page should be OK to edit again. --walter

Push Singh Memorial Fund Information

A memorial service for Push took place at the MIT Chapel at 3:00pm Thursday, March 9th. memorial service message



Dr. Push Singh

The MIT community lost a dear friend and colleague on 28 February 2006. Push had been a member of the community since he came to MIT as an undergraduate (Class of 1998). He was currently a postdoc and was planning to join the faculty in September 2007. His passions were Barbara Barry, Marvin Minsky, his family, friends, and colleagues, and understanding how the human mind works. While perhaps best known for laying the foundation for giving computers human-like common sense--"the ability to think about the everyday world like people do"--his interests ranged from developing a theory of beauty to developing an architecture for reflective thinking. He was recently named one of the IEEE Intelligent Systems 10 to Watch.

Push had a smile that would melt you. He was a gentle, caring person, but deeply passionate and, while soft-spoken, he never shied away from penetrating, yet constructive, criticism. Like his mentor, Push was generous with ideas; he gave us many ideas "new to mankind." He is missed by his friends and will leave a hole in our personal and intellectual lives.

Collecting Push's Work
We would like to gather Push's many writings, the best email dialogues fit for publication, and other works for a posthumous publication. Please upload it to Push's Work.

There are some videos of Push defending his thesis and talking about his research plans and articulating his vision of the future of the mind.

Pictures




Shashi Kant
I knew Push for only a short time during our many discussions about the intersection of the field of the AI and Semantic Web. He wanted to make the OMCS based available in RDF and make it more usable. This is such a shock! His death is a great loss to MIT and the scientific world - a flower was snipped off in its prime. God bless your soul Push, may you rest in peace my friend.

Aaron Sloman
I met Push in 1995 and over many interactions at conferences, exchanging ideas by email, and just chatting on the rare occasions when we had the opportunity, grew to like and admire him enormously. Impressive as his achievements are, his potential was even greater.

Jeff Patmore
Push always took time out to chat with me when I visited the Media Lab something I really valued. His enthusiasm was infectious and his insight marvellous.

Eyal Amir
Push was a happy person, a bold researcher, and a kind friend. I loved him and will miss him a lot in all ways. He has made waves with his research, an imaginative and free spirit, and it will affect the world not to have him with us.

Bo Morgan
Here is a story of one way Push has changed my life that I read at the MIT memorial service: One story of how Push changed my life.

Push had the uncanny ability to show you the future, inspire and lead those around him to work toward independent and novel ways of thinking. Push was a teacher who led by guiding his student's momentum and inspiration toward recognizing the low hanging fruit.

A funny anecdote that I'd like to share: Push would often talk about robotics and the benefits of simulating robots rather than building the actual hardware because the more abstract high-level intelligence domains, such as social reasoning and life goals, could be dealt with in a simulated environment with multiple humans being simulated. As a Ph.D. student Push programmed a very realistic rigid-body physics simulation, which began as a simulation for Push's Baby as the UROPs jokingly referred to the creature that he was programming (while simultaneously building OpenMind). Over the years that Push worked on his Ph.D. this robot baby with a society of mind grew up into two teenage segway robots with critic architectures of mind that could interact socially. Just recently, Push was talking with Dustin and I, and commented, "You know all that work we did programming that physics simulation for years? None of it was actually AI." Push had a way of emphasizing the forefront and unexplored parts of the AI space, always taking your idea and showing you what it looks like from about 50 miles up. This is probably obvious from how he explained in his Ph.D. how programs can debug themselves within the fun details of the social robot simulation, but it's funny how critical he was of the space that he was exploring, even the parts of the space that he himself had charted.

Joanna Bryson
I met Push when he was UROP for Rod Brooks on Cog in its first year. I think he'd already done UROPs with Pattie Maes and Marvin Minsky, and he was reputed to have read every book on all three of their bookshelves during the year he worked for each of them. He was one of the few people I ever knew at the AI lab who really could and wanted to talk about AI, everything about AI.

Marvin was the head of my OQE committee so my exam was over at the Media Lab. Lots of people speculate about what happens while they're out of the room during their OQE, but since the Media lab was all glass, you could actually watch my committee get up and shout at each other. Push noticed me waiting and came over and talked to me the whole time they were arguing, which was not only kind but really way more interesting than the exam.

Can we only talk about the good things here? Feel free to edit me if so. I didn't understand why it was getting harder to talk to Push about AI, why the free exchange of ideas was getting less and less free, although he was still very friendly and I remember going to dinner with him and Barbara in the early noughts. I didn't even get to see him though the last time I was in town. I wondered why he wasn't answering mail, but I didn't pursue it. One of the last email exchanges I had with him was an argument, he insisted that AI was sick and making no progress as a field, which I thought was crazy (though the progress may be different than what we expected, every year we can do more, so that can only be good.)

But that never made sense to me, and I still can't really remember it. When I think about him, I think about sitting in the sunshine on the lawn in front of the bio building debating with immense enthusiasim whether common-sense reasoning required fully grounded, validated understanding or could be constructed from rote memorization. Or about meeting him and Barbara in the Celler Bar on Mass Ave. Or eating Vietnamese food in Montreal. Well, all over the place but always engaged, always enthused --- the setting didn't matter to him --- well, it was fun --- but what mattered were the ideas.

Matt Weber
Push was my boss for a pseudo-UROP (I wasn't an MIT student) at the Media Lab in 2000, when OMCS was just starting up. I still haven't come down on how I feel about the project, in large part because it's changed so drastically since I worked on it, but Push was always way nicer to me than my competence merited and more liberal with department-sponsored lunches than he probably had any right to be, and my discussions with him and my experience on OMCS definitely helped engage me with the cognitive sciences, where I still am and ever hope to be. Bo's story has reminded me of another -- one day, while nothing in particular was happening, Push leaped from his computer, wide-eyed, and ran out the door to the elevator. We were all completely mystified; I, at least, thought something terrible had happened. As it turned out: While Push and Bo had been trying to hack his physics simulator to get a robot to walk, Push had gotten really interested in the baby of a colleague, who was just on the cusp of walking. He'd sprinted from the lab because the baby had just taken her first steps -- at the Media Lab, no less.

Matt Gorbet
I met Push in late May 1995 when I was a UROP working at the Media Lab. I remember the date because it was the week that Java (called HotJava) had been announced, and he walked up as I was looking at the first demos of the technology - we sat silently staring at a grey-and-black line drawing of an analog clock that was actually ticking away and showing the correct time in a web browser. Whenever I think of Push I think of that moment I shared with him when the WWW changed dramatically. I also remember the next hour or so, when he and I chatted - he spoke softly and asked me tons of questions about myself and life in general. I'll never forget that about him: he was always asking questions and always deeply interested in what seemed like the most mundane things. I was startled at first by how deeply he pursued everything I said, as though he actually cared. It turned out that he did care, deeply, about people's stories and what makes people tick.

Rajiv Rawat
This story is a bit different as it precedes and does not overlap Push's career at MIT. I used to say when introducing Push to my friends that he knew me before I was born. Our families were neighbours in Dehradun, India and next door neighbours in Canada from the first time we arrived in this country in 1975, to 1990, when I moved away, and a year later, Push moved to Cambridge. Eventually, I would join him in Boston from 1997 to 2002, although I would rarely get to see him as he was so immersed in his work.

We were each other's oldest friends, and acted much more like brothers. He being older would lead with an interest, whether it be computer programming (beginning with typing in games on a VIC-20, and saving to a cassette tape drive, all the way to the Amiga), adventure game playing (on an Apple IIe), Dungeons and Dragons, science fiction novels, or a million other things that cluttered our nerdy childhood. We gazed through telescopes, sat in our makeshift tree house, built snow burroughs in the front yard (can't do this anymore due to global warming), lay in the camper for hours that was parked in the driveway, launched model rockets, created indestructably dense lego spaceships, and watched "V", the 1983 miniseries more than a dozen times (although I never got to see Mike Donovan pulling the mask off the Reptilian Alien, as I'd always run and hide. Push tried different things to get me to watch but I always escaped). Going further back, we used to mimic Indian movie actors, where he'd always choose Amitabh Bhachan and me, Shashi Kapoor. Another game we played was "Nitroglycerin" where the floor of his unfinished basement meant instant death, sending us clambering over storage boxes stacked along the sides of the room.

By high school, I used to come over and walk to school with him (although he always cut it really close, so started walking on my own), and we'd bike over to the computer store. Sometimes, we'd go downtown to the science fiction bookstore. He helped me with my seventh grade science fair project (an egg shaped, triple jointed Martian) called "webbo" that won first prize. This was part of a bigger fascination for life out in the universe that began with Star Trek and extended to Benford, Brin, Varley, Herbert, and many other science fiction authors which he introduced me to over the years.

By the end of high school though, our lives had diverged. Already he had discovered his path by reading and analysing the "Society of the Mind", while I was getting involved in progressive activism and an eclectic series of pursuits. We did have dinner together once in Boston at one of his favourite places, Kebab'n'Kurry on Mass Ave, but otherwise I saw less and less of him throughout the 1990s. The last time I talked with Push was wishing him happy birthday last year, that came to my mind out of the blue. Sadly, I would have contacted him soon to get his address for a wedding invitation.

This might sound corny, but I felt at the funeral that they should play Amazing Grace. Spock's death scene in Star Trek II, where Kirk eulogized him as being the most "human" being he had ever met in his travels, seem appropriate to Push, who was at once intellectually curious and logical (or as he put it, sensible) and deeply human.

Words can't express how big a loss this is, and while we celebrate his life, we should also grieve his passing. He was and is loved so deeply.

Bradley Rhodes
I think the last time I saw Push was when I visited the Lab in June 2004. The satellite offices over at 1CC had been vacated, and Push and a few other late-term grad students had essentially set up shop with computers and whatever furniture they could find. I remember joking with him that he was already acting like faculty, right down to knowing how to scrounge for resources...

David G. Stork
Around 1997 I started to realize the opportunities of collecting data contributed over the internet and started what would become the Open Mind Initiative. At that time, there was noone doing quite what I envisioned. Then I heard that there was a graduate student at MIT interested in this subject too, specifically for the collection of commonsense data. When I finally got in touch with Push, we shared a deep sense of commaraderie -- we both were very excited about these new opportunities and he was bursting with ideas on how to pursue them. It was clear he would be exploring this area in great depth -- as indeed he has.

Over the years, through my visits to MIT, his visits to my lab, rendezvous at conferences and workshops and through innumerable emails, I came to know him as kind, throughtful, very clever and productive. His work I knew was first-rate, though occasionally tinged with ambition that might have humbled others. His Open Mind Commonsense project will be remembered in A.I. books and the research literature for a long time, I predict.

I was stunned to learn of his death. It was profoundly sad to think that he had a faculty position waiting for him at the Media Lab that he would never enjoy, a covetted position that he rightly earned. We can only imagine what contributions he would have continued to make to our field.

His friends, colleagues, family, MIT and the broader world of A.I., will mourn his loss -- a loss made so poignant by his youth, his promise, and his engaged and generous character.

Roy Rodenstein
I was a grad student in Sociable Media from 1998-2000, and my office was just a couple of doors down from Push's so I spent many a late-night hour talking with Push and being fascinated by his baby project. I had taken Marvin's class and was always interested in AI, so talking with him was always a treat. Without exception, he was always approachable, friendly, calm. The last time I saw him was at the Lab's 20th Reunion in October. If anything, he seemed happier than ever. It's a great loss.

Bob Hearn
I had the privilege of knowing Push for seven years. When I arrived at MIT, wanting to build the Society of Mind, Push was there with many invaluable suggestions and perspectives. Later, as I struggled with nailing down specific architectures for polynemes, pronomes, etc., Push was always the perfect person to bounce ideas off of. They always came back better. Of all the grad students who overlapped with me, Push's goals were closest to my own in trying to understand and build minds from a Society of Mind perspective. I looked forward to a long and fruitful future of working together. It is hard to believe he is gone.

Marvin and Gloria Minsky
This is the worst of times and so sad because it seemed to have been the best of times. Push was like a comet lighting up the intellectual sky with his brilliant, deep ideas and with his beautiful personality--and then moving into the darkness of this unexpected loss of his life.

None of us will ever adapt to this loss to Push's family and mine, to Barbara, and to the students he worked with and taught--and to the many researchers with whom he developed so many new and important ideas. To all of these he also served as a model of intellectual power, kindness, and honesty.

Still, his blazing trail remains in his many writings, publications, notes, and in the memory networks of his friends. We were all just beginning to know the range and the depth of his ideas. Now, although we shall have to struggle without his presence, we do have many records with explanations of his plans for future projects.

Push had just been appointed to become a new professor at MIT, to pursue what we knew would be a brilliantly productive career. For several years we have been designing an ambitious project to develop those theories, which have slowly been gaining the interest of many researchers inside and outside of the Media Lab. To us his loss is indescribable because of how we could communicate so much and so quickly in so very few words, as though we were parts of a single mind. We will never forget our years of talking and teaching together about what might underlie the resourcefulness of human thought.

There will be a fellowship in his honor for undergraduate students at MIT. We also plan to review and assemble his writings into a website and a reference book for other researchers--so that his deep and unique ideas about commonsense and computers will light the way for future students.

Henry Lieberman
The loss of Push Singh is at once a terrible personal tragedy for me and a tremendous loss for science.

In recent years, I had been enjoying a close collaboration with Push around the ideas of capturing and using Common Sense in AI. My group is called Software Agents and we work towards putting AI techniques to use in making user interfaces hopefully a little less stupid and frustrating than they are now. I thought that maybe we could improve things if there were some way, to, well... give computers a little common sense. Push's dedication to doing just that, attracted and fascinated me.

Push and his students worked on collecting Common Sense and developing architectures and tools surrounding the knowledge. My students and I concentrated on applying this knowledge to improve all kinds of interactive applications: browsers, editors, games, phones, etc. Push helped me teach my course, where we taught students how to understand and work with Common Sense knowledge. It has now grown to the point that there are numerous projects involving Common Sense in many, if not most, of the other groups at the Media Lab, and also elsewhere. This collaboration has been one of the most productive and fruitful of my career, and I thus owe him a tremendous debt that I shall never have the opportunity to repay.

I must admit, that when Push first launched Open Mind Common Sense back in 2001, though sympathetic, I was more than a little skeptical. Who would have thought that simply putting up a site asking people to type in Commonsense knowledge in English would ever amount to anything? And what could you possibly do with such a mess as might result? But Push's can-do attitude, "well, let's try it and see", won me over. I was constantly inspired by his ability to take seriously the idea that we could actually make progress even in areas that seemed impossibly vague or thorny. More than that -- we HAD to.

Push was, in some sense, "old skool" in the world of AI, and having been in AI a long time, this was one thing I really appreciated about him. By "old skool" I mean that I saw in Push a great example of the pioneer spirit of the earlier days of AI. This means not being afraid to tackle the big questions -- Common Sense, knowledge, problem solving, story understanding, vision; you know, the big questions.

Push's thesis work on EM-One was in every sense a ground-breaking model of the mind, one that gives life to Marvin Minsky's seminal Emotion Machine theory. EM-One could catalyze a resurgence of broad-spectrum models of thinking analogous to the way Terry Winograd's Shrdlu did when I was at the MIT AI Lab in the 1970s.

Today, we often get too fascinated with the latest hardware gadget or statistical trick. We need to remember that computers are really just a tool for understanding ourselves. That was Push's quest.

Walter Bender
From where did his vision emanate? Commonsense at MIT? Until his memorial service in Montreal, it had been an enigma to me that Push could infuse the cold field of AI with something so human. But hearing his family and friends tell their stories of Push put it all into context for me: they gave him a gift--his warmth--and he had begun a transformation of that gift into a model of mind that was attuned to human goals—a gift to all of us.

Ken Haase
I remember:


 * Push's slight smile, as he dug deep behind the deceptive ease of thought, challenging "what everybody knows" by simply making and building the pieces of minds to come.


 * Some holiday dinners at our home, where Push was always interested and interesting, and conversations ranged widely and warmly.


 * Push's easy grace, open heart, and the way he loved questions as much as answers, never afraid to grasp an idea and spin out possibilities and explanations to the horizon.


 * The fun of thinking about thinking with Push, playfully framing crystal whys and messy hows, probing mind and world with thoughts and code, founding dreams of wise machines to come.


 * The humble way he bridged individuals and projects who should really be talking with one another, but were separated by wide and sometimes bitter rivers.

Rather than dwell on how Push died, I choose to remember how Push lived, with imagination burning bright and energy rushing forward. Remembering that energetic light and warmth of mind and heart, I am both inspired to continue the work we shared and reminded to cherish the heart's joys and hopes alongside the mind's mysteries and answers.

I miss you Push. Like so many others on this page and beyond, I am greater for having known you and lesser for having lost you. May your spirit and memory endure in joy.

Steve Pollock
I knew Push since grade 2. I can't believe he is gone!

I will never forget you old pal.

Charlie Kemp
I have a strong memory of talking with Push outside of Prof. Minsky's office more than a decade ago, after I had first become interested in AI. Push impressed me then as one of the few people who was willing to consider the big picture of AI and directly confront the challenges of general machine intelligence. He was passionate, insightful, and kind. Over the years, my appreciation only increased as he began to deliver on his ambitions, and as I came to recognize the rarity of his perspective and talent. I am grateful for the time I was able to spend with Push, and I am deeply saddened that he is gone. It is a tremendous loss.

Junia Coutinho Anacleto and the Brazilian Open Mind group
I met Push last October, when I was at MIT to show our work in collecting and analysing Brazilian common sense. I feel that I am lucky having the chance of developing Push Singh 's OMCS Project in Brazil. It is being really exciting to work with a so brilliant idea. I thank him for being so generous with us.

(From student Fabiano Pinatti:)

Hi Barbara, Hugo and Henry, I’ve decided to write because I imagine you are suffering a lot from what happened to Push – Junia told me about the incident. I’m so sorry for that. I lost my family (mother, father, sister and a cousin) in a car accident in 2003 and I know how painful is to lose a beloved person. I know you will never forget him, but keep with you the good moments and continue your lives like he was with you. I’m sure he wouldn’t like to see you sad. Regards, Fabiano Pinatti

Benjamin Mako Hill
I met Push just over one year ago and do not know him as well as many of the others who have written their reminisces on this page. However, in only a short period of time and a limited number of interactions with Push, I came to be able to verify nearly every word of praise I have heard. Push was kind, gentle, caring, and compassionate. He was brilliant. Each quality was obvious to anyone who crossed his path.

Push once described something to me as advanced. When I pressed him for details on what he meant by this, he explained that by "advanced," he meant "ideas new to this earth."

Push was atypical in that he pursued advanced techniques and goals in artificial intelligence by trying to think well beyond what everyyone else was doing. He once told me that our job at the Media Lab was to be at least 10-15 years beyond the state of the art. Push was unusual in that he understood what this meant. He was extraordinary in that he consistently succeeded in realizing the future ahead of its time. Even at the Media Lab, Push lived up to the hype.

In the short time he was on this earth, Push brought this planet ideas, techniques, and methods that were truly advanced -- even when measured against his own impossibly high standards. Push's friends, MIT, the Media Lab, the AI community, and the rest of the earth will advance more slowly without him.

Wendy Ju
Push was my officemate when I first came to the Media Lab, and we became apartment mates at 1010 Mass Ave #47. I frankly had a hard time adjusting to MIT, so it was wonderful to be able to bookend each day talking to Push. He was always so philosophical and funny about the various events and misadventures of each day. It made it easier to gain some distance from my problems, to think about them in the abstract, as problems of reasoning or emotion.

My fondest memories with Push are of adventures we had on the way to MIT. One morning, when we were having bagels at the ABPi, we watched a dog walk the wrong way around a pole, oblivious to the fact that its owner had gone the other side--with the leash handle. We both laughed at the dog's utter surprise when his collar yanked him back. "This is the reason why cats don't let people leash them," I said. Push was really excited about this, and formed a theory about how cats understand string and dogs don't.

It was wonderful, really, to see how Push's work permeated his thoughts and his life. While his AI robot was learning to walk, Push and I would try to come up with a new way of walking to the T each day, to do some move that we were sure we had never done before. This made for some very goofy moments on Mass Ave. as we wriggled and shimmied our way to Harvard Square.

People are always surprised to hear that I knew Push before he started dating Barbara. "Before Push and Barbara!" they exclaim, "When was that!" Well, yes, before Barbara, Push dated a series of nice girls with very beautiful shiny hair. I enjoyed meeting them when he would invite them over to cook dinner with him. I distinctly remember, though, when Push started dating Barbara. They went ice skating on one of their first dates, and afterwards he kept mentioning her, saying, "I think there's something there" and "You would like her a lot." Push was absolutely right. I loved Barbara as soon as we met--and not just because she was nice, or had such beautiful, shiny hair. I loved Barbara so much so that I tried not to talk too much about Push, lest I give away that Push would leave dishes in the sink, or never bought more than one roll of toilet paper at a time. I was thrilled when they started dating seriously. They are my favorite couple still.

It's impossible to fathom this difference in our lives, with Push and without. It hurts to think about what might have been. I am trying hard to be philosophical, instead of upset, to be grateful, instead of angry. I am trying something new every day.

Ian Eslick
Here are some of the things I wanted, but was unable, to share at the memorial today:

For the past three years, Push Singh has been my colleague and mentor, but most importantly he was a very generous friend.

The first thing that to me defined Push was that he was first and foremost a builder and tinkerer. When he had an idea, he would go build something to try it out: construct a prototype, assemble a dataset or write down a giant lists of parts. Never was the his love and talent in creation clearer to me than when I first looked over some of his source code. The code simply screamed elegance. The layout was tight and clean, the comments clear and concise, and the abstractions amazingly consistent. There was the kind of attention to detail that you expect from a fine craftsman or an artist. After knowing Push a little longer I realized that this wasn't the painful labor of love it would have been for me, this was just how he thought.

I was a great admirer of Push's, though I probably didn't show it enough. He was marvelously stubborn. From taking on the harder thesis topic, to constantly bucking convention he did what he thought was best. I could sit here all week relating the wonderful arguments we had. Push never adopted an idea that he didn't make entirely his own. Time and time again we'd argue about something and then I'd hear nothing for a week, but sure enough, he would come back with a variation, extension or rebutt. The amazing thing was that whatever we talked about, he would continue to quietly think about it. You can't say that many people take others' ideas as seriously as he did. But despite his open nature, in the end, he always made his own way. I've come to associate this attribute as one shared by the best of all the minds I've known.

If Push was stubborn with his peers, he was unfailingly generous to his students. I have yet to meet a student who did not enjoy their experience working with him. He was always available, affable, good at listening and wonderful in channelling their excitement. Push was best at asking "what would happen if..." and inspiring a student to go find the answer. The project that he is most recognized for, Open Mind, started in this way - he argued about the idea with others, eventually made the idea his own and inspired students to run away with it. That idea is still running and will continue for some time. There are dozens of papers and even a workshop devoted to this pioneering idea.

On a more personal note. In the darkest of times there were two topics that pre-occupied Push during our discussions. The first and in his words the most important, was the great love of his life: Barbara. The goal he included me in, to construct a mind-like software system, was second. Of course he loved his family too and it showed best when we would go out to eat, he almost always had to eat Indian food and it was never as good as it was at home. He always would say that so you could tell what a big deal his home was and how much the food of his childhood meant to him.

The last thing I would like people to remember is his incredible passion to understand everything, especially the mind. I can't count the number of times I received a call from Push at 10 at night who would then launch into a discussion of his latest idea:

- What could we do if we collected all the goals people have?

- Could we put all of our commonsense onto a chip and embed intellgence into every cell phone?

- What happens if we put 1000 commonsense agents into a giant simulation and they all talk to each other?

- Hey, we should make a Commonsense Dating Agent that would know all the things a good nerd should not do on their first date. Maybe there's a company there?

I think it was this passion for the subject, an always playful personality that along with his strong sense of self, that allowed him to attempt many things that at the beginning were unfashionable, unpopular and often seemingly insane. But if some of the things he attempted and accomplished were controversial, it is only a testament to his bravery and the single minded pursuit of his own vision.

Much of what he thought, and more importantly how he thought, has indelibely altered my own ways to think. Should any idea of my own have impact in the future, Push will be an important part of it. I think that keeping his thoughts and ideas alive is the greatest gift any of us can give to one for whom life was defined by a passionate search for great ideas.

So go out right now and think of at least one insane idea.