Assistant Professor (Research)
Department of Cognitive, Linguistic & Psychological Sciences Brown University
Email | Google Scholar | GitHub | LinkedIn | CV
I develop AI systems inspired by biological vision and use advanced AI to understand the brain. My research bridges artificial intelligence and neuroscience to solve fundamental problems in both fields.
Research Focus
🤖 AI & Computer Vision
- Recurrent neural networks for visual reasoning
- Object tracking and perceptual grouping
- Biologically-inspired deep learning architectures
- Adversarial robustness and model interpretability
🧠 Computational Neuroscience & Biology
- Neural circuit discovery and reconstruction
- Brain-inspired algorithms for AI
- Visual perception and attention mechanisms
- AI-driven drug discovery and medical diagnostics
My work addresses a core question: Can we build better AI by understanding how biological brains process visual information? And conversely, can advanced AI help us decode neural circuits?
Previously: Postdoc with Thomas Serre (Brown University) | PhD, Boston College | BA, Hamilton College
Funding: National Science Foundation | Office of Naval Research
Research
Building Intelligent Machines that See Like Humans
My research program tackles two complementary questions:
- Can biological vision inspire better AI? I develop recurrent neural networks and attention mechanisms inspired by brain circuits to solve challenging computer vision tasks like object tracking, segmentation, and visual reasoning.
- Can AI help us understand the brain? I apply state-of-the-art deep learning to analyze neural data, reconstruct brain circuits, and discover new insights about biological vision systems.
Key Research Areas
⟳ Recurrent Vision Models
Building deep networks with feedback connections that mirror the brain’s visual cortex. These models excel at tasks requiring iterative reasoning, like contour detection and perceptual grouping.
Representative Work:
- Horizontal Gated Recurrent Units (hGRU) for long-range spatial dependencies
- Recurrent circuits for contour detection
- Stable and expressive recurrent architectures
👀 Visual Attention & Tracking
Understanding what humans look at and why—then building AI that does the same. Our ClickMe dataset contains half a million human attention maps for object recognition.
Representative Work:
- Tracking without re-recognition in humans and machines
- Learning what and where to attend
- Phase synchrony for appearance-changing object tracking
🤝 Biological-AI Alignment
Investigating why modern DNNs are becoming LESS aligned with human vision as they get more accurate. This work reveals fundamental differences between biological and artificial intelligence.
Representative Work:
- Performance-optimized DNNs evolving into worse models of visual cortex
- Harmonizing recognition strategies of humans and machines
- Better AI does not mean better models of biology
🔬 AI for Neuroscience & Medicine
Applying deep learning to biological problems: neural circuit reconstruction, drug discovery, cell death detection, and medical diagnosis.
Representative Work:
- Superhuman cell death detection for ALS/FTD research
- Deep learning for prostate cancer histopathology
- Neural scaling laws for phenotypic drug discovery
- Electron microscopy circuit reconstruction
Current Directions
I’m currently investigating:
- Visual perspective taking in humans and machines (3D-PC benchmark)
- Ecological data and objectives for human-AI alignment
- Neural mechanisms of visual simulation and mental imagery
- Computational demands of visual reasoning
Teaching: Computational Vision, Scientific Programming Organizing: Beyond Deep Learning Workshop Series Coordinating: Computation in Brain and Mind Initiative Datathons
Selected Publications
Google Scholar | View All Publications
2025
Better artificial intelligence does not mean better models of biology Drew Linsley, Pinyuan Feng, Thomas Serre arXiv 2025
2024
The 3D-PC: A benchmark for visual perspective taking in humans and machines Drew Linsley, Peisen Zhou, Alekh Karkada Ashok, Akash Nagaraj, Gaurav Gaonkar, Francis E. Lewis, Zygmunt Pizlo, Thomas Serre arXiv 2024 (accepted to ICLR 2025)
Tracking objects that change in appearance with phase synchrony Sabine Muzellec, Drew Linsley, Alekh Karkada Ashok, Ennio Mingolla, Girik Malik, Rufin VanRullen, Thomas Serre arXiv 2024 (accepted to ICLR 2025)
Deceptive learning in histopathology Sahar Shahamatdar, Daryoush Saeed-Vafa, Drew Linsley, Farah Khalil, Katherine Lovinger, Lester Li, Howard T. McLeod, Sohini Ramachandran, Thomas Serre arXiv 2024
Building better models of biological vision by searching for more ecological data diets and learning objectives Drew Linsley, Akash Nagaraj, Alekh Ashok, Francis Lewis, Peisen Zhou, Thomas Serre arXiv 2024
2023
Performance-optimized deep neural networks are evolving into worse models of inferotemporal visual cortex Drew Linsley, Ivan F. Rodriguez Rodriguez, Thomas Fel, Michael Arcaro, Saloni Sharma, Margaret Livingstone, Thomas Serre NeurIPS 2023
Unlocking feature visualization for deep networks with magnitude constrained optimization Thomas Fel, Thibaut Boissin, Victor Boutin, Agustin Picard, Paul Novello, Julien Colin, Drew Linsley, Tom Rousseau, Remi Cadene, Lore Goetschalckx, Laurent Gardes, Thomas Serre NeurIPS 2023
2022
Harmonizing the object recognition strategies of deep neural networks with humans Thomas Fel, Ivan Felipe Rodriguez, Drew Linsley, Thomas Serre NeurIPS 2022
Understanding the computational demands underlying visual reasoning Mohit Vaishnav, Remi Cadene, Andrea Alamia, Drew Linsley, Rufin VanRullen, Thomas Serre Neural Computation 34(5): 1075-1099
How and what to learn: Taxonomizing self-supervised learning for 3D action recognition Amor Ben Tanfous, Aimen Zerroug, Drew Linsley, Thomas Serre WACV 2022
2021
Development of a deep learning algorithm for the histopathologic diagnosis and Gleason grading of prostate cancer biopsies Ohad Kott, Drew Linsley, Ali Amin, Andreas Karagounis, Carleen Jeffers, Dragan Golijanin, Thomas Serre, Boris Gershman Clinical Genitourinary Cancer 2021
Superhuman cell death detection with biomarker-optimized neural networks Jeremy W. Linsley, Drew A. Linsley, Josh Lamstein, Gennadi Ryan, Kevan Shah, Nicholas A. Castello, Viral Oza, Jaslin Kalra, Shijie Wang, Zachary Tokuno, Ashkan Javaherian, Thomas Serre, Steven Finkbeiner Science Advances 2021
Tracking without re-recognition in humans and machines Drew Linsley, Girik Malik, Junkyung Kim, Lakshmi Govindarajan, Ennio Mingolla, Thomas Serre NeurIPS 2021
2020
Recurrent neural circuits for contour detection Drew Linsley, Junkyung Kim, Alekh Ashok, Thomas Serre ICLR 2020
Stable and expressive recurrent vision models Drew Linsley, Alekh Karkada Ashok, Lakshmi Narasimhan Govindarajan, Rex Liu, Thomas Serre NeurIPS 2020 Spotlight
2019
Small-molecule modulation of TDP-43 recruitment to stress granules prevents persistent TDP-43 accumulation in ALS/FTD Mark Y. Fang, Sebastian Markmiller, Anthony Q. Vu, Ashkan Javaherian, William E. Dowdle, Philippe Jolivet, Paul J. Bushway, Nicholas A. Castello, Ashmita Baral, Michelle Y. Chan, Jeremy W. Linsley, Drew Linsley, Mark Mercola, Steven Finkbeiner, Eric Lecuyer, Joseph W. Lewcock, Gene W. Yeo Neuron 2019
Disentangling neural mechanisms for perceptual grouping Junkyung Kim, Drew Linsley, Kalpit Thakkar, Thomas Serre ICLR 2020 Spotlight
2018
Learning long-range spatial dependencies with horizontal gated recurrent units Drew Linsley, Junkyung Kim, Vijay Veerabadran, Charles Windolf, Thomas Serre NeurIPS 2018
Learning what and where to attend Drew Linsley, Dan Shiebler, Sven Eberhardt, Thomas Serre ICLR 2019
Complementary surrounds explain diverse contextual phenomena across visual modalities David A. Mély, Drew Linsley, Thomas Serre Psychological Review 2018
2017
What are the visual features underlying human versus machine vision? Drew Linsley, Sven Eberhardt, Tarun Sharma, Pankaj Gupta, Thomas Serre ICCV 2017 Workshops
Resources
Code & Datasets
Recurrent Vision Models
- C-RBP Code – Recurrent vision models (2020)
- Recurrent Circuits Code – Contour detection (2019)
- hGRU TensorFlow – Horizontal gated recurrent units
- hGRU PyTorch – PyTorch implementation
Attention & Interactive Tools
- ClickMe Dataset – Half a million human attention maps
Perceptual Grouping & Pathfinding
- Clustered ABC Datasets – Easy, Medium, Hard variants
- Pathfinder Datasets – Versions 6, 9, and 14
Additional Resources
- Contextual Circuit Code – Visual phenomena modeling
Workshops & Teaching
Beyond Deep Learning Workshop Series Annual workshop exploring biological intelligence and AI at leading conferences.
Computational Vision Course Graduate-level course at Brown University covering modern computer vision and neuroscience.
CBMM Datathons Coordinating datathons through the Computation in Brain and Mind Initiative.
Contact
Get in Touch
Email: drew_linsley@brown.edu
Office: Department of Cognitive, Linguistic & Psychological Sciences Brown University Providence, RI 02912
Links:
Prospective Students
I’m always interested in working with motivated students passionate about the intersection of AI and neuroscience. If you’re interested in:
- Biologically-inspired deep learning
- Visual attention and reasoning
- Human-AI alignment
- Computational neuroscience
Please reach out with your CV and research interests.