EUVIMEDEUVIMED entdecken
Open Access

Inductive biases of neural network modularity in spatial navigation

Ruiyi Zhang iD, Xaq Pitkow iD, Dora E Angelaki iD

Science Advances · 2024

DOI: 10.1126/sciadv.adk1256 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Science Advances, 2024. Verfasst von Ruiyi Zhang und 2 weiteren Autor:innen. Beteiligte Einrichtung: Tandon School of Engineering, New York University, New York, NY, USA. Hauptthemen laut MEDLINE: Spatial Navigation, Neural Networks, Computer. Die Arbeit wurde bisher 12-mal zitiert (Crossref) und nennt 60 Literaturangaben. Der Volltext ist frei in PubMed Central verfügbar. Gefördert von: NINDS NIH HHS.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

The brain may have evolved a modular architecture for daily tasks, with circuits featuring functionally specialized modules that match the task structure. We hypothesize that this architecture enables better learning and generalization than architectures with less specialized modules. To test this, we trained reinforcement learning agents with various neural architectures on a naturalistic navigation task. We found that the modular agent, with an architecture that segregates computations of state representation, value, and action into specialized modules, achieved better learning and generalization. Its learned state representation combines prediction and observation, weighted by their relative uncertainty, akin to recursive Bayesian estimation. This agent’s behavior also resembles macaques’ behavior more closely. Our results shed light on the possible rationale for the brain’s modularity and suggest that artificial systems can use this insight from neuroscience to improve learning and generalization in natural tasks.

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

Spatial NavigationNeural Networks, ComputerAnimalsBayes TheoremNerve NetHumansBrainModels, NeurologicalLearning

Autor:innen und Institutionen

  • Tandon School of Engineering, New York University, New York, NY, USA.
  • Neuroscience Institute, Carnegie Mellon University, Pittsburgh, PA, USA.
  • Department of Machine Learning, Carnegie Mellon University, Pittsburgh, PA, USA.
  • Department of Neuroscience, Baylor College of Medicine, Houston, TX, USA.
  • Department of Electrical and Computer Engineering, Rice University, Houston, TX, USA.
  • Center for Neuroscience and Artificial Intelligence, Baylor College of Medicine, Houston, TX, USA.

ORCID-Profile: Ruiyi Zhang, Xaq Pitkow, Dora E Angelaki

Förderung

  • NINDS NIH HHS

Zitieren

APA

Zhang, R., Pitkow, X., & Angelaki, D. E. (2024). Inductive biases of neural network modularity in spatial navigation. Science Advances. https://doi.org/10.1126/sciadv.adk1256

Vancouver

Zhang R, Pitkow X, Angelaki DE. Inductive biases of neural network modularity in spatial navigation. Science Advances. 2024. doi:10.1126/sciadv.adk1256

BibTeX
@article{Zhang2024,
  author = {Zhang, Ruiyi and Pitkow, Xaq and Angelaki, Dora E},
  title = {Inductive biases of neural network modularity in spatial navigation},
  journal = {Science Advances},
  year = {2024},
  doi = {10.1126/sciadv.adk1256}
}
Kostenlos starten →