I recently got back from Portugal, where I was attending the Computational and Systems Neuroscience (CoSyNe) conference. This was my second time at the conference, having gone two years ago in Salt Lake City, but it was the first for which I was presenting research done in graduate school. I had an amazing time this year, and want to document and reflect upon some of the experiences that made this conference great (or have room for improvement).
Major themes of #cosyne19
As mentioned in the opening remarks of the conference, this year saw an unprecedented number of applications for poster presentations. Part of this was driven by increased interest from European researchers who in past years have not wanted to fly all the way to Salt Lake City. The other part, I suspect, was driven by increasing interest in AI and, by extension, computational neuroscience. Given that AI bigshot Yann Lecun was the keynote speaker, the conference organizers do not appear to be overly concerned by people associating Cosyne with AI. I don’t mind this per se, but the organizers need to plan for larger conference sizes if they wish for it to be integrated with the exploding AI community. It will be interesting to see whether the number of applications still increases despite moving back to Denver.
The other big demographic concern was related to the representation of women at Cosyne. Attendees were not satisfied by a wave at a bar graph showing that the number of women accepted to the conference were “roughly proportionate” to the number of female applicants. This is, after all, a conference of data nerds. And given the disproportionately small number of female selected speakers, the conference clearly needed to grapple with representation problems. Fortunately, the executive committee were quick to adopt measures proposed at a diversity lunch during the conference, such as double-blind review process and appointing a diversity & inclusion chair for next year. Seems like these should have been the default, but better late than never!
As for the actual talks, they were generally good. Janelia continues to pump out some of the most mind-blowing research around. I mean, something called the FlyPez that dispenses fruit flies one at a time like candy wasn’t even the coolest part of a talk by Gwyneth Card. To avoid conference burn-out, I mostly attended sessions featuring topics I am generally interested in, so there were a number that I missed. However, I saw enough to notice something of a trend. Many of the high-profile senior PIs that I wanted to see gave profoundly mediocre talks, which were either difficult to follow or boring surveys of previous research. Meanwhile, many talks by younger researchers were very engaging and told compelling stories about cutting-edge science. I’m not sure what, if anything, can be done about this, but perhaps a criterion for inviting a senior speaker should be the quality of talks they give.
Finally, it was interesting to see the pervasiveness of the word “manifold” in the conference. It’s my understanding that debate over whether this word was at all useful to describe the low-dimensional structure of neural activity in certain contexts reached a fever pitch at last year’s conference, and it looks like it’s become an established part of the lexicon. It will be interesting to see what words become important to computational neuroscience understanding in future years.
As I mentioned, I didn’t want to sit in talks all day, lest I lose my ability to focus and lost interest at the poster sessions. I also really wanted to explore the beautiful city of Lisbon. It was a very comfortable temperature during the day, and a mix of high-quality public transportation and cheap rideshares made it very easy to get around. There were also a ton of electric scooters, but they didn’t seem to do well on the city’s hilly streets. I had a great time exploring the Baixa, Alfama, and Belem neighborhoods. I cannot recommend the local pastry, pastel de nata, enough. The best place I found them was at a popular bakery in Belem, where the exact recipe is a carefully guarded secret. I also discovered the magic of fado, a special kind of Portuguese music that is meant to elicit strong feelings of saudade, which can be loosely translated as nostalgia.
After the talks were over, there was a shuttle to the nearby town of Cascais for the workshops. I was most interested in the “Data, Dynamics, and Computation” workshop on the first day, and attended all of the talks as well as a discussion at the end. The speaker line-up was great, and I enjoyed all of the talks. The discussion was also an interesting component. Maneesh Sahani acted as a facilitator, and did an admirable job keeping things moving. I’m generally a fan of open-forum discussions, but this was a large crowd (at least 100 people). It was intimidating to stand up and speak to the room, as I found out when I made a comment. Going back to the issues of representation, I noticed that there was only one woman who spoke during the entire discussion (though there were many in the room). It was also pointed out to me by a friend that the format is not friendly to people who speak English as a second language, since they may not be able to put their thoughts into words as quickly. I would love to see similar formats, but perhaps with smaller groups focused on particular sub-topics, in the future.
Cascais was a very small town with a beautiful waterfront, which I walked around on the first day. There was also a beautiful bike path along the coast to Guincho Beach. I did that on the second day, and saw very few people on the way. It was too cold to go swimming in the ocean, but perfect for biking and walking on the beach. The town also had some great places to eat. The seafood at Marisco na Praca was incredible. I also loved the restaurant called House of Wonders, which served delicious vegetarian food in a rooftop seating area.
Things to look forward to at #cosyne2020
As I understand it, Cosyne will be back in the Rockies next year, in Denver. I will probably go skiing for the second time in my life (the first being at the conference in 2017). I am also excited to see the ever-expanding network of people I know who are motivated by similar questions as me.
One thing I would like to see next year is something like a social media panel. Twitter has come to play a large role in the computational neuroscience community, and is a primary way for me to find out about interesting research and current debates within the field. I recognized many of my “Twitter heroes” at the conference but felt weird about approaching them. I think it could be very useful to have something like that. If anyone who reads this agrees (or has other cool ideas) feel free to reach out!