Arthur W. Juliani, PhD
Postdoctoral Researcher. Microsoft Research, New York City, NY, USA.

I am a postdoctoral researcher currently at Microsoft Research in New York City. My interests reside at the intersection of psychology, neuroscience, machine learning, and philosophy.
My graduate research focused on human navigation and computational models of the brain systems which support it. I used a combination of behavioral experiments and novel computational modeling to develop a deep learning inspired model of the hippocampal place cell system.
While at Unity Technologies, I was one of the founding developers of the Unity ML Agents toolkit. I also lead a project to develop and host the Obstacle Tower Challenge, a reinforcement learning competition in which hundreds of researchers benchmarked their algorithms.
During my time at Araya Inc I led a pair of projects studying the potential for current deep learning architectures to capture the functional properties of current leading models of consciousness. This included a proposal that the Transformer architecture can be understood as functional Global Neuronal Workspace.
At Microsoft Research I am currently focusing on using theoretical frameworks from deep reinforcement learning to model both healthy and pathological representation learning, belief formation, and decision making in humans. The goal of this line of research is to better understanding mental health disorders and their potential treatment through novel modalities such as psychedelics, meditation, and neurofeedback.
latest posts
Oct 4, 2023 | A Gentle Introduction to the Free Energy Principle |
---|---|
Jul 10, 2023 | AI Accelerationism is a Humanism |
Jun 14, 2023 | On The Promise and Paradox of Non-Psychedelic Psychedelics |
selected publications
2023
- Deep CANALs: A Deep Learning Approach to Refining the Canalization Theory of Psychopathology2023
2022
- On the link between conscious function and general intelligence in humans and machinesTransactions on Machine Learning Research, 2022
2019
- Obstacle tower: A generalization challenge in vision, control, and planningarXiv preprint arXiv:1902.01378, 2019