Blake Richards is a neurobiologist, Fellow in CIFAR’s Learning in Machines & Brains program, and an Assistant Professor at the University of Toronto, Scarborough Department of Biological Sciences.
What does it mean to be a CIFAR fellow?
Critically, what it means to me is to be tapped into a network of researchers who have similar interests in this sort of interdisciplinary research that links computer science with neuroscience. That’s not something you always find.
Being able to attend these meetings where I can interact with all these AI folks lets me gain influence from and think about the question of how the brain works from a perspective that I don’t get at standard neuroscience meetings.
What makes CIFAR program meetings different?
If you go to traditional neuroscience meetings, most people are biologists. And that’s not a bad thing. But if someone is purely a biologist, it means that they don’t have the kind of perspective on how you’d actually mathematically capture the complexity of the brain and how you’d come up with algorithms that actually lead to animal or human intelligence.
I would guess that other CIFAR programs would provide a similar opportunity to view your problem, the thing you study, in something other than the light that you get at standard conferences.
How does your CIFAR program contribute to your research?
In addition to being influenced by people in the program, it was in part the interactions at other CIFAR meetings that helped shape some of my thinking about how to approach some new ideas and even more so with guiding the follow ups.
The kind of interactions that you get from the Learning in Machines & Brains program meetings are fantastic for giving an opportunity to discuss these ideas between both neuroscientists and artificial intelligence researchers.
The other thing is a bit of money, and flexible money. In being a fellow, I have a little bit more breathing space in terms of my funding, like whether I take on a student for a unique project that doesn’t fit into my standard grants, and also the knowledge that if there is something that we really want to explore we can.
For example, there are some research questions that myself and some colleagues like Joel Zylberberg, who is a CIFAR Azrieli Global Scholar, are interested in. He and I met at a CIFAR meeting and ended up chatting a lot and came up with some ideas for how to experimentally measure the learning algorithms of the brain. We wanted to get a grant for this work but it’s difficult. It’s a hard sell with a lot of standard granting agencies where interdisciplinary work is not in line with standard dogma. CIFAR provided us with some catalyst grant funds.
In general, you have that sort of ability to go to CIFAR and say “Look, we’ve got this far-out idea. Can you give us a bit of support to just get the preliminary data to then apply to these other institutions?” It’s very helpful.
What was surprising, unexpected or different about being a CIFAR Fellow?
It was surprising to me just how many other people share my kind of interest in bringing together the AI theory with neuroscience. This is what’s so fantastic about the program.
The other thing that is surprising is in the way CIFAR structures its reporting requirements. It doesn’t inundate us with administrative work. We submit our yearly reports and it’s not a massive exercise in paperwork, which was a pleasant surprise.