Armin Thomas

[LinkedIn] [Google Scholar] [GitHub]

I am a Ram and Vijay Shriram Data Science Fellow at Stanford University, where I work in artificial intelligence research and its application to computational neuroscience and biology together with Russell A. Poldrack and Christopher Ré. At Stanford, I am also affiliated with Stanford Data Science and Stanford's Centers for Research on Foundation Models and for Open and Reproducible Science.

Prior to Stanford, I was a research fellow and mentor of the Max Planck School of Cognition, completed a PhD in AI/ML at Technical University of Berlin in the group of Klaus-Robert Müller, and worked as a research scientist in Antonio Rangel’s Neuroeconomics laboratory at Caltech (in collaboration with Google ATAP).

My research is focused on diverse research problems in between AI and computational neuroscience. Some of my recent research focuses are…

…training AI systems at scale (e.g., models trained on large-scale brain data [Paper]; genomic foundation models such as Evo (7B parameters) [Paper] and HyenaDNA (1M context) [Paper]; language models with up to 7B parameters [Paper] [Paper])

…advancing the ability of AI systems to learn from long sequences (e.g., [Paper], [Paper], [Paper], [Paper])

…developing and evaluating explainable AI tools for neuroscience research (e.g., [Paper], [Paper], [Paper])

…building computational models to better understand the algorithms underlying human choice behavior (e.g., [Paper], [Paper], [Paper])


* indicates equal contribution


Mechanistic Design and Scaling of Hybrid Architectures. Michael Poli*, Armin W. Thomas*, Eric Nguyen*, Pragaash Ponnusamy, Björn Deiseroth, Kristian Kersting, Taiji Suzuki, Brian Hie, Stefano Ermon, Christopher Ré, Ce Zhang, Stefano Massaroli. ArXiv. [Paper] [Code]

Sequence modeling and design from molecular to genome scale with Evo. Eric Nguyen*, Michael Poli*, Matthew G. Durrant*, Armin W. Thomas, Brian Kang, Jeremy Sullivan, Madelena Y. Ng, Ashley Lewis, Aman Patel, Aaron Lou, Stefano Ermon, Stephen A. Baccus, Tina Hernandez-Boussard, Christopher Ré, Patrick D. Hsu*, Brian L Hie*. bioRxiv. [Paper] [Code].


Monarch Mixer: A Simple Sub-Quadratic GEMM-Based Architecture. Daniel Y. Fu, Simran Arora*, Jessica Grogan*, Isys Johnson*, Sabri Eyuboglu*, Armin W. Thomas*, Benjamin Spector, Michael Poli, Atri Rudra, Christopher Ré. Advances in Neural Information Processing Systems (NeurIPS). [Paper] [Code]. Oral.

HyenaDNA: Long-Range Genomic Sequence Modeling at Single Nucleotide Resolution. Eric Nguyen*, Michael Poli*, Marjan Faizi*, Armin W. Thomas, Callum Birch-Sykes, Michael Wornow, Aman Patel, Clayton Rabideau, Stefano Massaroli, Yoshua Bengio, Stefano Ermon, Stephen A. Baccus, Chris Ré. Advances in Neural Information Processing Systems (NeurIPS). [Paper] [Code]. Spotlight.

Benchmarking explanation methods for mental state decoding with deep learning models. Armin W. Thomas, Christopher Ré, Russell A. Poldrack. NeuroImage. [Paper] [Code]

Evaluating deep transfer learning for whole-brain cognitive decoding. Armin W. Thomas, Ullman Lindenberger, Wojciech Samek, Klaus-Robert Müller. Journal of the Franklin Institute. [Paper] [Code].

Simple Hardware-Efficient Long Convolutions for Sequence Modeling. Daniel Y. Fu*, Elliot Epstein*, Eric Nguyen, Armin W. Thomas, Michael Zhang, Tri Dao, Atri Rudra, Christopher Ré. International Conference on Machine Learning (ICML). [Paper] [Code].

Hungry Hungry Hippos: Towards Language Modeling with State Space Models. Tri Dao*, Daniel Y. Fu,* Khaled K. Saab, Armin W. Thomas, Atri Rudra, Christopher Ré. International Conference on Learning Representations (ICLR). Preprint: [Paper] [Code]. Spotlight.


Interpreting mental state decoding with deep learning models. Armin W. Thomas, Christopher Ré, Russell A. Poldrack. Trends in Cognitive Sciences. [Paper].

Self-supervised learning of brain dynamics from broad neuroimaging data. Armin W. Thomas, Christopher Ré, Russell A. Poldrack Advances in Neural Information Processing Systems (NeurIPS). [Paper] [Code].

Differentiable programming for functional connectomics. Rastko Ciric, Armin W. Thomas, Oscar Esteban, Russell A. Poldrack. Machine Learning for Health Workshop at NeurIPS. [Paper] [Code]. Best Poster.

Gaze-dependent evidence accumulation predicts multi-alternative risky choice behaviour. Felix Molter, Armin W. Thomas, Scott A. Huettel., Hauke R. Heekeren, Peter N. Mohr. PLoS computational biology. [Paper] [Code].


On the opportunities and risks of foundation models. Rishi Bommasani, R., …, Armin W. Thomas, ... & Percy Liang. ArXiv. [Paper].

Uncovering the computational mechanisms underlying many-alternative choice. Armin W. Thomas, Felix Molter, Ian Krajbich. Elife. [Paper] [Code].


Machine learning methods for modeling gaze allocation in simple choice behavior and functional neuroimaging data on the level of the individual. Armin W. Thomas. Technische Universität Berlin, Berlin. [Paper].


Gaze bias differences capture individual choice behaviour. Armin W. Thomas*, Felix Molter*, Ian Krajbich, Hauke R. Heekeren, Peter N. Mohr. Nature human behaviour. [Paper] [Code].

Analyzing Neuroimaging Data Through Recurrent Deep Learning Models. Armin W. Thomas, Hauke R. Heekeren, Klaus-Robert Müller, Wojciech Samek. Frontiers in Neuroscience. [Paper].

GLAMbox: A Python toolbox for investigating the association between gaze allocation and decision behaviour. Felix Molter*, Armin W. Thomas*, Hauke R. Heekeren, Peter N. Mohr. PloS one. [Paper] [Code].

Deep transfer learning for whole-brain FMRI analyses. Armin W. Thomas, KLaus-Robert Müller, Wojciech Samek. Machine Learning in Clinical Neuroimaging Workshop at MICCAI 2019. [Paper].


I believe in open science and therefore put strong emphasis on open sourcing all code and data used for my research and teaching. Find some examples of open source projects below:


Teaching & Tutorials:

Thanks to my mentors 🙏🏻

I’ve been very fortunate to have been advised by brilliant mentors, among them Russell A. Poldrack (Stanford), Christopher Ré (Stanford), Hauke R. Heekeren (Hamburg University), Antonio Rangel (Caltech), Klaus-Robert Müller (TU Berlin), Ian Krajbich (UCLA), and Ulman Lindenberger (Max Planck). I’m grateful to them for their advice and support through the years.

📩 gmail: athms.research