DJ Strouse

I'm a Research Scientist at DeepMind in London.

I did a PhD in Physics at Princeton University, advised by David Schwab and Bill Bialek, and funded by a Hertz Fellowship and a DOE Computational Science Graduate Fellowship. I did a master's at the University of Cambridge with Mate Lengyel as a Churchill Scholar and studied physics and mathematics at the University of Southern California, where I worked with Bartlett Mel and Paolo Zanardi. During my studies, I also spent time at DeepMind working with Matt Botvinick, Stanford University working with Kwabena Boahen, the Institute for Quantum Computing working with Andrew Childs, and Spotify NYC working with their machine learning team.

Email  /  Twitter  /  CV  /  Blog  /  Google Scholar  /  GitHub

News
  • I accepted a job at DeepMind beginning in March 2019.
  • Our paper on promoting generalization and exploration in RL by training agents with information bottlenecks between goal and policy was accepted to ICLR 2019.
  • I presented our paper on promoting cooperation / competition via information regularization at NeurIPS 2018.
  • I defended my PhD in August 2018.
Research

I'm interested in reinforcement learning (RL), information theory, and deep learning with an eye toward understanding and creating intelligent agents. Much of my work has focused on the information bottleneck (IB) and lately I've been interested in using IB to improve the training of RL agents. In the past, I've also worked on quantum information theory and computational neuroscience.

Transfer and Exploration via the Information Bottleneck
Anirudh Goyal, Riashat Islam, DJ Strouse, Zafarali Ahmed, Hugo Larochelle, Matthew Botvinick, Sergey Levine, Yoshua Bengio
International Conference on Learning Representations (ICLR), 2019
OpenReview / my DeepMind talk / Matt's NIPS talk

We train agents in multi-goal environments with an information bottleneck between their goal and policy. This encourages agents to develop useful "habits" that generalize across goals. We identify the states where agents must deviate from their habits to solve a task as "decision states" and show that they are useful targets for an exploration bonus.

Intrinsic social motivation via causal influence in multi-agent RL
Natasha Jaques, Angeliki Lazaridou, Edward Hughes, Caglar Gulcehre, Pedro A. Ortega, DJ Strouse, Joel Z. Leibo, Nando de Freitas
NIPS Emergent Communication Workshop, 2018
arxiv / OpenReview

We reward agents for influencing the actions of other agents, and show that this gives rise to better cooperation and more meaningful emergent communication protocols.

Learning to share and hide intentions using information regularization
DJ Strouse, Max Kleiman-Weiner, Josh Tenenbaum, Matt Botvinick, David Schwab
Neural Information Processing Systems (NIPS), 2018
arxiv / nips / code / bibtex

We train agents to cooperate / compete by regularizing the reward-relevant information they share with other agents, enabling agents trained alone to nevertheless perform well in a multi-agent setting.

Efficient use of discrete latent variables using the deterministic information bottleneck
DJ Strouse, David Schwab
in progress
note

We use the variational deterministic information bottleneck to regularize discrete latent variable models, encouraging the use of as few latent variables as possible for a given level of performance.

The information bottleneck and geometric clustering
DJ Strouse, David Schwab
forthcoming in Neural Computation (NECO)
arxiv / code / bibtex

We show how to use the (deterministic) information bottleneck to perform geometric clustering, introducing a novel information-theoretic model selection criterion.

The deterministic information bottleneck
DJ Strouse, David Schwab
Neural Computation (NECO), 2017 & Uncertainty in Artificial Intelligence (UAI), 2016
arxiv / code / UAI / NECO / bibtex

We introduce the deterministic information bottleneck (DIB), an alternative formulation of the information bottleneck that uses entropy instead of mutual information to measure compression. This results in a hard clustering algorithm with a built-in preference for using fewer clusters.

Using neural networks to understand the computational role of dendrites
DJ Strouse, Balazs Ujfalussy, Mate Lengyel
Computational and Systems Neuroscience (Cosyne), 2012 & 2013
2012 poster & abstract / 2013 poster & abstract / master's thesis / why

We fit neural network models to single neuron data to understand the computational role of dendrites in integrating their synaptic input.

Behaviorally-locked structure in a sensory neural code
DJ Strouse, Jakob Macke, Roman Shusterman, Dima Rinberg, Elad Schneidman
Sensory Coding & the Natural Environments (SCNE), 2012
abstract / poster

We study the olfactory neural code in mice and find that much of the information about the stimulus is only decodable when interpreting neural activity relative to the sniff phase, providing evidence for the importance of considering sensory sampling behavior when interpreting neural codes.

Optimizing online learning capacity in a biologically-inspired neural network
Xundong Wu, DJ Strouse, Bartlett Mel
Society for Neuroscience (SfN), 2011 & Computational and Systems Neuroscience (Cosyne), 2012
SfN poster / Cosyne abstract

We study the optimal conditions for online recognition memory in a biologically-inspired neural network with "dendrite-aware" learning rules.

Levinson's theorem for graphs
Andrew Childs, DJ Strouse
Journal of Mathematical Physics (JMP), 2011
arxiv / JMP / talk

We prove an analog of a classic result in quantum scattering theory for the setting of scattering on graphs. The goal is to provide additional tools for designing quantum algorithms in this setting.


Good artists copy.