Simone Azeglio

Simone Azeglio

PhD Candidate at the Vision Institute & École Normale Supérieure

Key Interests: Representation Learning, Vision, Biological Inspiration & Inductive Bias, Equivariant Representations, Generative Models, Information Theory, Alignment

Hello! I’m currently a PhD Candidate at Sorbonne University’s Vision Institute and École Normale Supérieure in Paris, under the supervision of Ulisse Ferrari, Peter Neri and Olivier Marre.

My research bridges computational neuroscience and machine learning by formalizing key information processing principles from biological vision into three complementary frameworks:

  1. Statistical principles of higher-order correlations in natural scenes and their processing by visual circuits, enabling more expressive feature extraction beyond standard convolutional approaches

  2. Geometric principles of visual processing, particularly equivariance to transformations like scale and velocity, which provide powerful inductive biases for robust perception

  3. Information-theoretic analysis and manifold structure of visual representations, enabling precise decomposition of stimulus features that contribute to perception and neural encoding

Previously, I worked as a Research Engineer in Brice Bathellier’s Lab at the Hearing Institute (Institut Pasteur), developing activity-driven frameworks for the auditory pathway. I also spent time as a research intern at the Flatiron Institute (Simons Foundation) working on Bayesian inference and deep learning models for image and video encoding with Alex Williams, and gained experience on ML at CERN (with Lorenzo Moneta) and on nonlinear models of spatial memory at the University of Ottawa with André Longtin and Leonard Maler.

I’ve organized workshops at major conferences including the “Symmetry and Geometry in Neural Representations” workshop at NeurIPS (2022-2023-2024), the “Symmetry, Invariance and Neural Representations Workshop” at the Bernstein Conference (2022-2023), and the “Sharpening Our Sight Workshop” at Cosyne 2024.

I’m also the co-founder and member (ex-president) of the Machine Learning Journal Club (MLJC), a non-profit research organization focusing on interdisciplinary applications of Machine Learning (Scientific ML, Brain-Computer Interfaces and Natural Language Processing). Please feel free to contact me if you’re interested in any of these projects or share similar research interests.

News

Happy to Open NeurReps 2024 in Vancouver!

Join Us for NeurReps @ NeurIPS 2024

The 6th Episode of our Seminar Series it taking place @ U of Amsterdam

The 5th Episode of our Seminar Series it taking place @ Harvard

The 4th Episode of our Seminar Series it taking place @ MIT

The 3rd Episode of our Seminar Series it taking place @ UPenn

The 2nd Episode of our Seminar Series it taking place @ MILA

The NeurReps 2024 CfP is out!

Andy Keller is our 1st Speaker

NeurReps WorldWide Seminar Serie Announced

=”utf-8”></script>

SoS workshop @Cosyne2024 Confirmed Speaker’s List

SoS workshop @Cosyne2024 is taking place

NeurReps @NeurIPS 2023 is taking place!

NeurReps submission deadline gets extended!

Group Theory in Neuroscience @ Symmetry, Invariance and Neural Representations

Symmetry, Invariance and Neural Representations is happening again @ Bernstein 2023

Art and AI’s Video Representations @ Sorbonne

NeurReps Proceedings are out 🎉

It took a while, but finally the proceedings of our (with Sophia Sanborn, Christian Shewmake, Arianna Di Bernardo and Nina Miolane) workshop are out 🎉

3rd Place in Sensorium Competition @NeurIPS 2022

Together with my supervisors: Ulisse Ferrari, Peter Neri and Olivier Marre, I took part in the Sensorium 2022 competition on predicting mice V1 (Primary Visual Cortex) neural activity with deep learning methods. Our team, “IdV_ENS” got 3rd and I had the chance to present our work at NeurIPS 2022 🐭🧠

NeurReps: Symmetry and Geometry in Neural Representations is LIVE @NeurIPS 2022

Glad that our (together with Sophia Sanborn, Christian Shewmake, Arianna Di Bernardo and Nina Miolan) workshop, Symmetry and Geometry in Neural Representations a.k.a. NeurReps has been a truly enjoyable occasion. If you’re interested in following further developments at the intersection of Neuroscience, Deep Learning and Geometry don’t hesitate to join our community on Slack

Our Workshop has been accepted @NeurIPS 2022

Our workshop on Symmetry and Geometry in Neural Representations has been accepted to @NeurIPSConf 2022! We’ve put together a lineup of incredible speakers and panelists from 🧠 neuroscience, 🤖 geometric deep learning, and 🌐 geometric statistics.

Brains for Brains Award 2022

Since 2010, the Bernstein Network Computational Neuroscience presents the Brains for Brains Award, which recognizes the special achievements of young scientists who have shown their outstanding potential at a very early career stage – even before starting their doctoral studies. It is one of two prizes awarded bi-yearly by the Bernstein Network. This award aims to attract young international researchers to Germany.

Symmetry, Invariance and Neural Representations

Azeglio & Di Bernardo, Symmetry, Invariance and Neural Representations Bernstein Conference 2022

Improving Neural Predictivity in the Visual Cortex with Gated Recurrent Connections

Azeglio, Poetto, Savant-Aira, Nurisso, Improving Neural Predictivity in the Visual Cortex with Gated Recurrent Connections Brainscore Workshop - Cosyne 2022