The success of CBMM and the people who made it possible
Date Posted:
January 10, 2025
Date Recorded:
January 10, 2025
CBMM Speaker(s):
Samuel Gershman ,
Josh McDermott ,
Carlos Ponce ,
Maria Fernanda De La Torre ,
Leyla Isik ,
Boris Katz ,
James DiCarlo ,
Kohitij Kar ,
Gabriel Kreiman ,
Winrich Freiwald ,
Mengmi Zhang ,
Will Xiao ,
Morgan Talbot ,
Ed Boyden ,
Akshay Rangamani
Description:
Over the last 11+ years, the Center for Brains, Minds, and Machines has become a place of gathering brilliant minds to discuss and solve the challenges and questions of intelligence, organic and artificial. Hear from a few of our members, and others, about the impact and people the Center has made.
[MUSIC PLAYING] JAMES DICARLO: My name is Jim DiCarlo. And I'm the director of the MIT Quest for Intelligence and the co-director of the Center for Brains, Minds, and Machines. In fall of 2023, we celebrated CBMM's remarkable history.
FERNANDA DE LA TORRE ROMO: I remember seeing the sign, Center for Brains, Minds, and Machines outside of Tommy's office and feeling this, like, joy of, like, oh, this is what I ever wanted. Like, I didn't know that a place that is thinking about all these things would exist.
SAMUEL GERSHMAN: Really good interdisciplinary interactions is not something you see every day in academia, even though it's a sort of value that is promulgated. And I think the CBMM did a really remarkable job bringing people together that were far enough apart that they could learn something from each other, but also close enough together so that they could effectively talk to each other.
LEYLA ISIK: It feels like this interdisciplinary area between brains, minds, and machines is so popular now. But I think a lot of that popularity is thanks in large part due to CBMM and the intellectual influence that's had on so many people in the field.
CARLOS PONCE: Some of the brightest minds in neuroscience and machine learning came together to discuss important topics, to tell each other what they're thinking about, why it is important, and in a way, inspire the rest of us, at the time, I was a trainee, to really think about what is it that we want to do in the field of neuroscience in general.
FERNANDA DE LA TORRE ROMO: And so that has been more challenging than I originally expected. But it's also been much more rewarding than I could have imagined, because it introduces you to a community of people that love talking about the mind, love talking about their own thought process, and getting to questions that are very curiosity driven, where you don't find that outside of this community. And so it's been more rewarding than I thought. Like, it's not just about the work, but the immersiveness of how everybody else is thinking.
BORIS KATZ: CBMM created a new field, the science of intelligence, which will be deeply important in the future.
SAMUEL GERSHMAN: CBMM has produced hundreds of research papers. But I think more fundamentally, it's changed the direction of how people think about important problems within the science of intelligence.
JOSH MCDERMOTT: It provided a way for this whole generation of students to kind of start getting together and talking to each other, and to start to learn to speak a common language, to get exposure to other faculty that kind of were in the same space.
KOHITIJ KAR: But this is the life that I got to live. And I was very happy living this life of cross-collaboration at CBMM.
LEYLA ISIK: And I think for a trainee that was just really, really unique to get to interact with all these giants in the field in this, like, interdisciplinary area of intelligence.
GABRIEL KREIMAN: It has provided a forum for people to dialogue, and brainstorm, and think about new projects and new ideas.
CARLOS PONCE: I enjoy that sense that we were part of a larger community, that we were part of a larger effort, and that it would take all of us coming together.
WINRICH FREIWALD: And that's really allowed us to tackle questions that otherwise we would not have been able to tackle.
JOSH MCDERMOTT: It's given the students exposure to this really kind of cutting-edge perspective. And so our students have learned about all the stuff that's kind of become very influential. But I think they learned about it before almost everybody else. And so they're kind ahead of the game. And they've become leaders in the field as a consequence. And I think that is really in large part because of CBMM.
KOHITIJ KAR: A lot of people from here will go found startup companies, will found labs that will be leading the field. And I think this is a testimony to the success of CBMM.
LEYLA ISIK: I feel like so many of the people I know in the field, I met through CBMM first. And it's just such a cool way to have spurred a lot of collaborations and friendships.
SAMUEL GERSHMAN: The legacy of CBMM will not just be in the research it produced, but also in all of the artifacts that are now publicly available for education.
CARLOS PONCE: I feel that we've arrived at a place where the questions are within our grasp. We can actually try and answer these questions, but we have to do it together. And this isn't going to be something where a single investigator, or a single team, or a single department can really tackle the complexities of these problems. And as technology gets better, as everything advances, we have more tools. We have more opportunities. We have more resources. And now, the question becomes, what are we going to do with these tools? How are we going to put them all together?
And this is the kind of thing that CBMM helped us all understand. By providing a larger umbrella where ideas can be exchanged in a practical way, that is going to be I think, important for the field in general. And that's where I think CBMM was instrumental.
JAMES DICARLO: Perhaps the most important legacy CBMM leaves is the Brains, Minds, and Machines summer course. Hundreds of students have come to the summer school, spending three weeks in an intensive learning environment. And the students bring varied perspectives and experiences that lead them to new ideas and research. Many students return as teaching assistants and then pursue academic careers and return as instructors. Others have entered industry or launched startups. And there's no greater success story for a school.
GABRIEL KREIMAN: The summer course was created at the beginning of CBMM with the intention of creating a new field, essentially, and training a new generation of students that could fluidly converse between neuroscience, and cognitive science, and AI.
MENGMI ZHANG: It's a very competitive program. They are selected among 300 candidates. We selected 30 plus students. They are amazing. They are from all over the world.
JOSH MCDERMOTT: It's a chance to spread the gospel, if you will, to kind of take this perspective that we think is really important here, and to give people from elsewhere an opportunity to hear a little bit about that.
BORIS KATZ: It is an opportunity to help hundreds of young researchers to change the trajectory of their research, and give them the knowledge to study intelligence in all its complexity.
CIANA DEVEAU: I have really enjoyed the summer course. This is actually the second time that I applied to be here. That's how much I was excited about it. It's been really great through the lectures and the tutorials to just get this expansive understanding of all different kinds of both models and cognitive science in the ways that we can kind of bring these modalities all together to try and understand intelligence.
STUDENT: We had a workshop about theory.
STUDENT: AI connected to human intelligence.
STUDENT: Neuroscience.
STUDENT: And computer science.
CIANA DEVEAU: And on robotics.
STUDENT: Deep learning, and about machine learning in general.
FERNANDA DE LA TORRE ROMO: It made professors feel very accessible, like I could just sit down next to them at lunch and have a conversation about life or about science. And that's something I hadn't experienced before.
BORIS KATZ: My participation in the summer school is one of the most meaningful and satisfying experiences in my life.
KOHITIJ KAR: I'm a faculty now, and I don't need to be here for three weeks spending my summer here. But I still choose to. And I think that at least for me, it tells me that this is an important factor in my life. It's kind of the best three weeks of the year academically sometimes for me. And one of the highlights of that is the interaction with the students.
BORIS KATZ: We will always need an influx of young people to build on our achievements and to take it into the future.
ED BOYDEN: We need people to be able to think about the brain and intelligence from every angle, from molecules, to circuits, to algorithms, and everything in between.
BORIS KATZ: And the best way to do it was to educate young researchers from all over the world to make them equally comfortable in neuroscience, cognitive science, and AI.
GABRIEL KREIMAN: So we think that it's important, in addition to traditional lectures and discussions, to have students learn by doing. And that involves specific projects, whereby students are working on new ideas, usually outside of their comfort zone.
MORGAN TALBOT: A fantastic opportunity to take a risk and dive into a field that you never tried before.
CIANA DEVEAU: My project is definitely outside of my comfort zone. I am learning a lot of new skills, which is really great.
MORGAN TALBOT: As time went on, people started kind of branching out and doing almost side projects within the summer course. There was one at one point to control a drone using EEG signals from the brain.
SANTOSH VEMPALA: Some of them said that time is passing so quickly, even though they've been here for over a week. In the past, the theory workshop used to be 2/3 of the way down. And even at that point, people were not fatigued. They were excited. They were engaged.
SABINE MUZELLEC: I've had the opportunity to meet so many people, had so many interesting discussions, learning so much also. And yeah, extremely motivating and stimulating environment for science.
CIANA DEVEAU: If I have a question, there's always somebody that I can go to to get a really great answer, and feedback, and just point me to really helpful resources.
SAMUEL GERSHMAN: Students have come back with a number of assets from these courses. One of them is just basic skills. But I think maybe the more important, and sort of more difficult to quantify asset that they've taken away, is really a kind of broader thinking about the science of intelligence.
JOSH MCDERMOTT: I mean, this was really the first course on the science of intelligence from this perspective. And it continues to be very influential.
CIANA DEVEAU: It has instilled maybe a bit more confidence for me for moving forward and being like, OK, I do understand these principles more deeply.
MORGAN TALBOT: It also really kickstarted the project that I think now will be the focus of my dissertation.
SABINE MUZELLEC: I had actually a friend who was who's telling me, should I apply? And after two days, I told him, you definitely should. It's really great.
JAMES DICARLO: Many of us can't think about CBMM without thinking of Tommy Poggio. Tommy has been a mentor to generations. And we are so grateful for the role he has played in shaping our field through his research, his academic advising, but most of all, his friendship and support.
GABRIEL KREIMAN: In the context of CBMM, Tommy was the main originator of the original grant proposal submitted to NSF.
BORIS KATZ: Tommy is the founder, the director, and the soul of CBMM.
JOSH MCDERMOTT: It's remarkable how effective he has been as a scientific leader in bringing together all these disparate branches of the science of intelligence. And I think it really took somebody with his level of experience, and seniority, and ambition. And that's kind of what it took to pull together something like CBMM.
SAMUEL GERSHMAN: And what I appreciated about his leadership was that even though he had his own strong ideas about particular research questions, he didn't impose that on the other researchers. And he was very willing to pursue a kind of pluralistic vision of intelligence through CBMM, and try to foster different interactions between views of intelligence.
GABRIEL KREIMAN: Sometimes people say, an afternoon in the library can save you a month of work. I think it's fair to say that one hour with Tommy can save you a decade or two of going in the wrong direction.
AKSHAY RANGAMANI: Every other week, you get a new draft, or a new download, of, oh, this is how deep learning works. You haven't had time to check his hypothesis on the other ones first, but Tommy just keeps going.
GABRIEL KREIMAN: Tommy has been quite influential in the field, not only in terms of the specific work that he's done, and shedding light into the basic mathematics of learning, and the theory of learning, but also pioneering computational models of visual object recognition, the biophysics of computation, and neurons, and many other fields.
CARLOS PONCE: And a lot of our experiments kind of gave results that I understood fit best with what had been laid out by Tommy Poggio and his collaborators 30 years ago. That intellectual legacy is everywhere now in neuroscience. And it really is helping us put together a lot of the discoveries that we have now with advanced technology in a larger context.
GABRIEL KREIMAN: It's hard to go anywhere in science without meeting alumni from Tommy. So he's trained generations of amazing people all over the world. And I keep discovering new people. I go to a conference, I have dinner with people, and then all of a sudden, I discover that at some point, they passed through Tommy's lab, they interacted with Tommy, they collaborated with Tommy, or they were his former students, or postdocs.
LEYLA ISIK: When I think about it now, the amount of both resources, and support, and also freedom he gave me as a student was really amazing, and I think really, really helped me grow as a scientist.
FERNANDA DE LA TORRE ROMO: I think from him, I learned a lot about how to find the right questions in science, and how to be critical of certain things.
LEYLA ISIK: He's a really fun person to do research with. He's a very fun person to talk about science with. But more generally, he's just a very fun person to talk with. And that always made it feel really nice to be in his lab and be working with him.
FERNANDA DE LA TORRE ROMO: And to me, he seems like a real truth seeker. I think he doesn't care so much about who's right, but rather what the truth is. And I think I really got to see that while I was working with him. And it was inspiring.
LEYLA ISIK: I think I actually would be remiss to talk about Tommy's many wonderful qualities, and not also mention the fantastic espresso he always had on hand, too. So--
CARLOS PONCE: The three words that I would use to describe Tommy are, first of all, brilliant, second of all, inspiring, and third, visionary.
LEYLA ISIK: OK, he's obviously brilliant and a very-- a visionary in the field.
FERNANDA DE LA TORRE ROMO: Visionary.
SANTOSH VEMPALA: Visionary.
SAMUEL GERSHMAN: I admire Tommy's ability to look far in the future and sort of see what's coming, and anticipate it.
MENGMI ZHANG: Pioneering.
WINRICH FREIWALD: Pioneer.
AKSHAY RANGAMANI: Imaginative.
SANTOSH VEMPALA: Fundamental.
WILL XIAO: Wise.
GABRIEL KREIMAN: Intelligence. I think he's probably one of the smartest people I have ever met.
FERNANDA DE LA TORRE ROMO: Inspirational.
MENGMI ZHANG: Influential.
WILL XIAO: Curious.
GABRIEL KREIMAN: Infinite curiosity. So he's really touched upon themes like mathematics all the way to neuroscience all the way to cognitive science, robotics, self-driving cars. It's hard to think about fields that he hasn't really impacted.
MENGMI ZHANG: Multidisciplinary.
WINRICH FREIWALD: Theoretician.
SAMUEL GERSHMAN: Tommy has a lot of grit in the way that he herded all the cats that needed to be herded in such a diverse group of people.
WILL XIAO: I think Tommy is really approachable--
AKSHAY RANGAMANI: Energetic.
SANTOSH VEMPALA: And friendly.
LEYLA ISIK: He's just a really fun collaborator to have.
GABRIEL KREIMAN: And I think he's just an amazing, good person, and human being, and honest, and nice.
WINRICH FREIWALD: And gentleman.
FERNANDA DE LA TORRE ROMO: Thoughtful.
LEYLA ISIK: He's a extremely supportive mentor.
BORIS KATZ: A scholar, a leader, and a friend.
AKSHAY RANGAMANI: Can I say chocolate sorbet? Because he's always looking for that wherever he goes.
[MUSIC PLAYING]
have an interactive transcript feature enabled, which appears below the video when playing. Viewers can search for keywords in the video or click on any word in the transcript to jump to that point in the video. When searching, a dark bar with white vertical lines appears below the video frame. Each white line is an occurrence of the searched term and can be clicked on to jump to that spot in the video.
