82 citations
- Norwegian University of Science and TechnologyNO14 papers
- Stockholm UniversitySE7 papers
- KTH Royal Institute of TechnologySE6 papers
- The Abdus Salam International Centre for Theoretical Physics (ICTP)IT6 papers
- Nordic Institute for Theoretical PhysicsSE4 papers
- Scuola Internazionale Superiore di Studi AvanzatiIT4 papers
- University of CopenhagenDK4 papers
- Institute for Advanced StudyUS3 papers
- Capital Fund Management (France)FR2 papers
- Istituto Nazionale di Fisica NucleareIT2 papers
- Istituto Nazionale di Fisica Nucleare, Sezione di TriesteIT2 papers
- Aalto UniversityFI1 paper
19 papers
Efficiency of local learning rules in threshold-linear associative networks
Francesca Schönsberg, Yasser Roudi, Alessandro Treves
We derive the Gardner storage capacity for associative networks of threshold linear units, and show that with Hebbian learning they can operate closer to such Gardner bound than bi…
Self Punishment and Reward Backfill for Deep Q-Learning
Mohammad Reza Bonyadi, Rui Wang, Maryam Ziaei
Reinforcement learning agents learn by encouraging behaviours which maximize their total reward, usually provided by the environment. In many environments, however, the reward is p…
Minimum Description Length codes are critical
Ryan John Cubero, Matteo Marsili, Yasser Roudi
In the Minimum Description Length (MDL) principle, learning from the data is equivalent to an optimal coding problem. We show that the codes that achieve optimal compression in MDL…
Statistical Criticality arises in Most Informative Representations
Ryan John Cubero, Junghyo Jo, Matteo Marsili +2
We show that statistical criticality, i.e. the occurrence of power law frequency distributions, arises in samples that are maximally informative about the underlying generating pro…
Decoding of neural data using cohomological feature extraction
Erik Rybakken, Nils Baas, Benjamin Dunn
We introduce a novel data-driven approach to discover and decode features in the neural code coming from large population neural recordings with minimal assumptions, using cohomolo…
The Stochastic complexity of spin models: Are pairwise models really simple?
Alberto Beretta, Claudia Battistin, Clélia de Mulatier +2
Models can be simple for different reasons: because they yield a simple and computationally efficient interpretation of a generic dataset (e.g. in terms of pairwise dependences) -…