João Sacramento

Research

I'm a principal investigator at the Institute of Neuroinformatics and a guest researcher at Angelika Steger's group in the Computer Science Department of ETH Zürich. My research interests lie at the intersection of neuroscience and machine learning. I try to understand the principles that enable humans to learn so efficiently, when compared to artificial neural networks. My goal is to develop better neuroscience-inspired machine learning algorithms, and in turn to use insights gained from designing these to understand learning in the brain.

My research is supported by a SNSF Ambizione Fellowship and an ETH Zurich Research Grant.

Previously, from 2015 to 2018, I was a postdoc with Walter Senn at the University of Bern. Together with Rui P. Costa and Yoshua Bengio we developed a model for error backpropagation in the cortex.

I received my PhD in computer science from IST (University of Lisbon, 2014) where I studied neural network models of memory with Andreas Wichert. Still at IST, in 2015, I was awarded a short research fellowship to work with Francisco C. Santos. During this period I studied energy-efficient synaptic plasticity rules with Mark van Rossum.

Recent papers

	Zucchet N*, Schug S*, von Oswald J*, Zhao D, Sacramento J (2021). A contrastive rule for meta-learning. Preprint: arXiv:2104.01677 [ paper ] * — equal contributions

 

	von Oswald J*, Zhao D*, Kobayashi S, Schug S, Caccia M, Zucchet N, Sacramento J (2021). Learning where to learn: Gradient sparsity in meta and continual learning. NeurIPS 2021 [ paper ] * — equal contributions

 

	Meulemans A*, Farinha MT*, Ordóñez JG, Aceituno PV, Sacramento J, Grewe BF (2021). Credit assignment in neural networks through deep feedback control. NeurIPS 2021 (Spotlight) [ paper ] * — equal contributions

 

	Henning C*, Cervera MR*, D'Angelo F, von Oswald J, Traber R, Ehret B, Kobayashi S, Grewe BF, Sacramento J (2021). Posterior meta-replay for continual learning. NeurIPS 2021 [ paper ] * — equal contributions

 

	Jordan J, Sacramento J, Wybo WAM, Petrovici MA*, Senn W* (2021). Learning Bayes-optimal dendritic opinion pooling. Preprint: arXiv:2104.13238 [ paper ] * — equal contributions

 

	von Oswald J*, Kobayashi S*, Sacramento J*, Meulemans A, Henning C, Grewe BF (2020). Neural networks with late-phase weights. ICLR 2021 [ paper | code ] * — equal contributions

 

	Zhao D, von Oswald J, Kobayashi S, Sacramento J, Grewe BF (2020). Meta-learning via hypernetworks. NeurIPS Workshop on Meta-Learning 2020 [ paper | poster ]

 

	Meulemans A, Carzaniga FS, Suykens JAK, Sacramento J, Grewe BF (2020). A theoretical framework for target propagation. NeurIPS 2020 (Spotlight) [ paper | code ]

 

	von Oswald J*, Henning C*, Sacramento J*, Grewe BF (2019). Continual learning with hypernetworks. ICLR 2020 (Spotlight) [ paper | talk video | code ] * — equal contributions

 

	Richards B*, Lillicrap TP*, ..., Sacramento J, ..., Therien D*, Körding KP* (2019). A deep learning framework for neuroscience. Nature Neuroscience 22:1761-1770 [ link to journal ] * — equal contributions

 

	Llera M, Sacramento J, Costa RP (2019). Computational roles of plastic probabilistic synapses. Current Opinion in Neurobiology 54:90-97 [ link to journal ]

 

	Sacramento J, Costa RP, Bengio Y, Senn W (2018). Dendritic cortical microcircuits approximate the backpropagation algorithm. NeurIPS 2018 (Oral) [ paper | talk video ]

If my articles are behind a paywall you can't get through please send me an e-mail.

A complete list of publications is here.

Teaching

Currently, I'm a guest lecturer for the Learning in Deep Artificial and Biological Neuronal Networks course offered at ETH Zürich.

In the past I served as a teaching assistant at the Department of Computer Science and Engineering of IST, where I lectured practical classes on computer programming and basic algorithms.