In this blog, EDIT Lab PhD student Tim writes about his experiences at his first international conference.
Tim, PhD student
I had planned to finish the diagrams during the plane journey. Then I would have had the legroom to code and run models later that night. Unfortunately, the restrictions on when you could and couldn’t use your laptop during the Wednesday morning Aer Lingus flight to Dublin couldn’t be circumvented as easily as the man next to me sneaking glances at his phone. My knees, pressed into the back of the seat in front, were now fully visible with my Mac in the overhead locker, whilst he tutted away at Twitter.
Instead, I made my diagrams on the bus from the airport to Trinity College Dublin, where the first Computational Psychiatry Conference was being held. (#CPC, not the other CPC held that year, though blue was the lanyard colour of choice for both). My usual dilapidated sense of direction allowed me to check out the walls of this citadel placed in the centre of Dublin, as I searched for the entrance tucked into brickwork, and more time to make the poster that surely everyone else already had printed, tubes in hand.
Ducking to avoid the low doorframes, I chucked my stuff in my allocated room. The bed was recessed into the wall, and bookended by cupboards. I knew I wouldn’t fit; but that was a future Tim problem. I caffeinated my Mac and myself and typed up more words and arranged pictures on PowerPoint in the shape of a poster. I let some models run on all four cores whilst I headed out to eat.
The next morning, I overslept. I missed the complimentary breakfast. No black or white pudding. The models failed overnight, but I had some graphs on the page. And some words about reinforcement learning. I spent the hour I’d planned to go running in, instead attempting to save LaTeX equations as vectors. This didn’t work; the model equations were instead just a pasted jpeg, which would surely pixelate blown up to A0. There was no time to align all the textboxes. There may or may not have been a full stop at the end of every sentence. Resigned, I emailed it to the printers, hours after their deadline, but before the pre-conference tutorial session, which was about to begin.
Fuelled only by protein bars, I joined the registration queue at the exact moment some of the organising committee, which was fortunate as I could then ascertain the local bakery scene. There was running to be done, and Irish starches to be consumed. Steph told me about cronuts in Bread 41, and I could think of little else all afternoon. But they are only available first thing in the morning, she added, so I instead focused on what was in front of me.
The first day, or afternoon, was an optional tutorial session. £10 gets you a series of talks on psychiatric diagnosis, machine learning, and a tutorial on reinforcement learning. Three hours into the scheduled two-hour Yael Niv sermon on the latter, the hours of life devoted to this topic in the past year felt reclaimed.
Best travelled poster award goes to…
I now had one hour to pick up my poster, fit in a 10k run, and make it back in time to Trinity for the opening reception, in the Old Library. This may have been the first time Dublin has seen a man running along the Liffey with a cardboard tube. If there was a prize for the best-traveled poster, this would surely podium.
The actual conference, from the opening networking reception Thursday evening, to the formal dinner Saturday night, was an intense blur of nourishing content and interesting ideas. I stood by my poster Friday afternoon, to present to passers-by. No-one mentioned its flaws; they seemed more interested in the content. The crowd was friendly and supportive, in a city which facilitated this collaborative spirit. They gave PhD and early career researchers platforms alongside accomplished PIs. We could stay talking in Doyles till 3 am, which is where I gleaned most of the analytical insights I now work on.
I was too late for the cronuts on both Friday and Saturday mornings. And in fact, they were not cronuts, but cruffins. Again, I could think of nothing else, and knew I would have to pounce on my final morning.
I had booked a late flight back Sunday, to make the most of this new city. I ran Phoenix Park, in now familiar slight drizzle. I went to art galleries and museums. Found the one pub which served Guinness 0.0%, with a very put out bartender given he had to travel to the basement to find it.
But no cruffins. Not on Sundays, they said.
My tube safely in the hold, with no surcharges through Aer Lingus’s superior customer service, I ambled through the airport. I read half a book during the plane journey back, the position of my knees and my Mac far from my mind. I pressed my head against the plane window when the cabin lights went off to land, and saw the places I run in London in miniature.
I often think back to my time in Dublin, and my time at the conference. I’ve not yet used the many ideas it generated within me. Leaving the cloister of free flowing inspiration, and returning to the slog of coding it all up, with Stan models failing to converge for no apparent reason. But as reinforcement learning proposes, we learn from repeated experiences. And there is always next year’s CPC. I hear Minneapolis does wonderful éclairs.
