Alexander Meulemans

Curious about how the world works, I started a bachelor in Engineering Sciences at KU Leuven in 2014. During these studies, I gravitated towards the mathematical beauty of modelling this world and creating new learning systems and that is why I started a master of Mathematical Engineering, also at KU Leuven. In my exchange and master thesis at ETH Zürich, I got excited about the synergy of neuroscience and artificial intelligence, leading me to start a PhD at the lab of prof. Angelika Steger in ETH Zürich. Through mathematics, machine learning and neuroscience I want to provide a creative and fresh view on the 'intelligence' part of artificial and human intelligence. On the one hand, I design new theories for credit assignment in the brain (how does a neuron know how to change its synapses to improve the global behavior?), which is crucial for understanding learning in the brain, but surprisingly also very useful for learning on next-generation (analog) hardware, which has similar locality constraints as the brain. On the other hand, I draw inspiration from human intelligence to investigate currently unsolved problems in AI, such as long-term credit assignment in reinforcement learning and active perception. Besides my love for science, I'm a passionate musician and on the weekends, you can find me either skiing or hiking in the beautiful Swiss Alps.

Publications

The least-control principle for learning at equilibrium
Alexander Meulemans*, Nicolas Zucchet*, Seijin Kobayashi*, Johannes von Oswald, João Sacramento,
arXiv preprint, 2022. 

Minimizing Control for Credit Assignment with Strong Feedback
Alexander Meulemans*, Matilde Tristany Farinha*, Maria R. Cervera*, João Sacramento, Benjamin F. Grewe,
ICML, 2022 (spotlight presentation). 

Credit assignment in neural networks through deep feedback control
Alexander Meulemans*, Matilde Tristany Farinha*, Javier Garcia Ordonez, Pau Vilimelis Aceituno, João Sacramento, Benjamin F. Grewe,
NeurIPS, 2021 (spotlight presentation). 

Challenges for Using Impact Regularizers to Avoid Negative Side Effects
David Lindner*, Kyle Matoba*,Alexander Meulemans*
SafeAI workshop - AAAI, 2021. 

Neural networks with late-phase weights
Johannes von Oswald*, Seijin Kobayashi*,Alexander Meulemans, Christian Henning, Benjamin F. Grewe, João Sacramento
ICLR, 2021. 

Continual Learning in Recurrent Neural Networks
Benjamin Ehret*, Christian Henning*, Maria R. Cervera*,Alexander Meulemans, Johannes von Oswald, Benjamin F. Grewe
ICLR, 2021. 

A theoretical framework for target propagation
Alexander Meulemans, Francesco S. Carzaniga, Johan A.K. Suykens, João Sacramento, Benjamin F. Grewe
NeurIPS, 2020 (spotlight presentation). 

Check Google scholar for an updated and complete publication history.