Research Jam

Research Jams are a presentation of local research, but solely for unfinished (or better yet: unstarted) research ideas. They don't require slides, or results, or a solid idea, but should have a fruitful seed for others to jam. The seed can be a crazy idea, an interesting dataset, or a previous research project.

One rule: No negative statements will be tolerated. Research Jams should encourage people to present ideas they wouldn't dare present anywhere else; without fear of humiliation or legal action.

All Research Jams will be located in the Decision Theater, UVM Trinity Campus, Farrell Hall (first floor), 210 Colchester Ave, Burlington VT 05405

Host: Laurent Hébert-Dufresne


Thursday, May 17
VCET conference room (downstairs in Farrell Hall).

Hugh Garavan, from the Department of Psychiatry

In this research jam, I will describe the ABCD study ( which is a longitudinal study of adolescent development with a very large sample and very rich characterization of participants (brain imaging, genetics, cognition, hormones, psychological testing, environment). I will canvass suggestions on a contest or hackathon that we might run to improve and popularize the analyses of this very large dataset.

Monday, April 9
Decision Theater, Farrell Hall

Randall Harp, from the Department of Philosophy, will be leading a jam session based on The conceptualisation of the field of ethics in data science and the ideas behind

Monday, February 12
Decision Theater, Farrell Hall

Laurent Hébert-Dufresne

We will jam around an interesting dataset of community behavior collected during the Ebola outbreak in Sierra Leone.

Monday, March 12
Decision Theater, Farrell Hall

Matt Mahoney
from the Department of Neurological Sciences, will be leading a jam session based on the following idea:

Epilepsy is often described as a “network” disease, but everyone’s network is different and nobody knows what to do about it. In particular, there are myriad mathematical models of seizures as nonlinear dynamical systems, but these models require manually setting ~30 biophysical and anatomical parameters that are known to vary substantially across patients. Ideally, we seek to learn something about these parameters on a per patient basis by observing their brain activity alone, as this could help us non-invasively localize their underlying brain abnormalities. In this Research Jam, I would like to explore the idea of using a dynamical system as a generative model for machine learning on time series data, and discuss possibilities for mining the clinical EEG data that are being collected at UVMMC.