#  Clémentine Dominé 

Postdoctoral Fellow

 

 

 



   ![Clem](/sites/g/files/omnuum6471/files/styles/hwp_4_5__320x400/public/2026-10/41d36bd5-f135-4982-90f2-0b858ca4a67e%203.jpg?itok=6UBNPd5U) 

 



 

 laptop\_windows [Personal Website](https://clementinedomine.github.io/) 

 laptop\_windows [Google Scholar](https://scholar.google.com/citations?user=oVZ0fSYAAAAJ#%20your%20Google%20Scholar%20ID) 

 

 



 

Clémentine Dominé is a CBS-NTT Postdoctoral Fellow in the Physics of Intelligence at Harvard University. Her research lies at the intersection of theoretical neuroscience and theoretical machine learning. Broadly, she aims to understand how biological and artificial neural networks learn, adapt, and build representations that support complex behaviour, including continual, curriculum, and reversal learning, as well as the acquisition of structured knowledge. She develops mathematical frameworks rooted in deep learning theory to study these processes and to understand when learned representations can be shared, transferred, or reused across tasks, with connections to transfer learning, fine-tuning, and linear probing in machine learning.

  
Previously, she was a postdoctoral researcher at ISTA, working with Marco Mondelli and Francesco Locatello, and completed her PhD at the Gatsby Computational Neuroscience Unit under the supervision of Andrew Saxe and Caswell Barry. She holds a degree in Theoretical Physics from the University of Manchester, including an exchange at UCLA. She also co-organizes the UniReps workshop at NeurIPS, which brings together researchers studying representations across biological and artificial neural systems.



 

 

 





 

 

- ## People
    
     [Current Members](/job-roles/current-group-members-0)