About me

I’m a Postdoctoral Research Fellow in the NeuroAI and Geometry Data Analysis Lab at Harvard University. I work on developing a geometric characterization of predictive representations in artificial and biological neural networks.

I completed my PhD in the Zenke Lab at the Friedrich Miescher Institute in Basel, Switzerland. I worked on understanding the mathematical principles behind predictive self-supervised learning in order to see if and how the brain could be learning in a similar way. In the past, I studied Electrical Engineering at the Indian Institute of Technology in Madras, and Robotics at the École polytechnique fédérale de Lausanne. I also spent a fun half year interning with the Robotics AI team in Amazon, Berlin.

Publications

Implicit variance regularization in non-contrastive SSL
Halvagal, M. S.*, Laborieux, A.*, & Zenke, F.
Advances in Neural Information Processing Systems 36 (NeurIPS 2023)

( paper | code )

The combination of Hebbian and predictive plasticity learns invariant object representations in deep sensory networks
Halvagal, M. S., & Zenke, F.
Nature Neuroscience

( paper | code )

Inferring subjective preferences on robot trajectories using EEG signals
Iwane, F., Halvagal, M. S., Iturrate, I., Batzianoulis, I., Chavarriaga, R., Billard, A., & Millán, J. D. R.
2019 9th International IEEE/EMBS Conference on Neural Engineering (NER)

( paper )

News

• Defended my PhD thesis on “Predictive Self-Supervised Learning in Brains and Machines” (April 2025)
• Presented a poster at Bernstein 2024
• Presented a talk and tutorial at the Janelia Junior Scientist Workshop for Theoretical Neuroscience 2023
• Presented a poster at NeurIPS 2023 (main conference)
• Presented a poster at the Self-Supervised Learning Workshop at NeurIPS 2022
• Presented a poster at Bernstein 2022
• Presented a poster at Cosyne 2022