Computational Interaction Summer School

Debrief of the 5th Computational Interaction Summer School in Columbia University, New-York

During summer 2019, I had the chance to participate to the 5th edition of the summer school on Computational Interaction, in Columbia University, New-York.

For the first time in my PhD, I had the chance to meet other researchers and students in HCI from another university. It was also a great pleasure to be hosted in this wonderful university. Here’s an informal summary of my experience here, illustrated with small comic strips (in french sometimes).

My first duty wheen I arrived

Computational Interaction is a sub-field of Human-Computer Interaction that which focuses on the modelling the interaction between a user and a system. The easiest way to have a clearer idea of what differs between HCI and Computational Interaction is to look at the definition in the book of references , in which most of the authors gave a talk at the summer school:

“Computational interaction applies computational thinking– thas is, abstraction, automation, and analysis– to explain and enhance the interaction between a user (or users) and a system. It is underpinned by modelling which admits formal reasonning and involves at least one of the following:

  • a way of updating a model with observed data from users;
  • an algorithmic element that, using a model, can directly synthetize or adapt the design;
  • a way of automating and instrumenting the modelling and design process;
  • the ability to simulate or synthetize elements of the expected user-system behavior”

The program

The entire week was organzed supervised by two local and friendly researchers, Brian A. Smith (Columbia) and Xiaojun Bi (Stony Brook University).

First day:

TODO

Second day:

TODO

Third day:

TODO

The funny and captivating talk of Prof. Nikola Banovic on inverse reinforcement learning for modeling human behaviors.

Fourth day:

TODO

Fifth day:

TODO

Last day: The Hackathon 🏅

The last 24 hours was dedicated to a Hackathon. We could chose a project based on one of the talk presented. With my friends Vishnu Nair and Yuheng Wang, we get inspiration from the talk of Jason Jacques and tried to build a Voice-Assisted Text Entry Corrections. The system would backed up a text-based autocorrect system with a voice-based override in case the text-based system did not provide the correct autocorrect result. The override would correct the text and change the text model so that future corrections would be affected. We attempted to use Flask and Tkinter to create a prototype system that consisted of a touchscreen keyboard with voice capabilities that connected to our custom-made model.

I really liked working with Vishnu and Yuheng, we really lent themselves to the game, and stayed working until 3 a.m. Eventhough the system was slow and wasn’t working that well, we were honored to win the second place via an audience vote 🎉

The winning team made a very impressive system in which a generative model was adapting artistic generations based on user rewards (using a reinforcement learning algorithm if I remember well).

Finishing 2nd was the cherry on top of a amazing week, and we could get a nice souvenir:

Our team and the two Summer School organnizers

Life in New-York

After this intense and and busy week, I had the chance to spend some time visiting NY with my new friends. Being a huge fan of live music, it was a blessing to be in NY. I really spent memorable moments with Vishnu, Alen, Feng, João, Harish, Gustavo among others. Hope to meet you again in conference or somewhere else.

From the rooftop of the impressive Empire Hotel, booked for the summer school participants. Right next to central park!
An french man in NY: Brooklyn Bridge and Jazz Big band at Columbus Circle!
Having fun in New-York with Vishnu, Feng, João and Alen! Hope to see you again