University of Cambridge > Talks.cam > Computational Neuroscience > Connectome-based Echo State Networks

  Browse A-Z

Log in

Cambridge users (raven)details Other usersdetails No account?details

Information on

Subscribing to talksdetails Finding a talkdetails Adding a talkdetails Disseminating talksdetails Help and Documentationdetails

Connectome-based Echo State Networks

Add to your list(s) Download to your calendar using vCal

If you have a question about this talk, please contact Daniel Kornai.

The Echo State Network (ESN) framework is an efficient recurrent neural network paradigm of importance for both neuroscience and machine learning. It is quick and efficient to train, and also has been suggested as a possible model of brain function. It is not fully known how network structure influences ESN functionality, dynamics, and robustness. We used biological networks to study this, compared with randomly initialised ESNs. In both biological and artificial neural network contexts, neurons and clusters demonstrate functional specificity. We asked if synapse-resolution connectome-based ESNs demonstrate functional specificity/generality at the neural level. We used the larval Drosophila melanogaster connectome, which exhibits a hierarchical modular structure according to type, class, and function. We built connectome-based ESNs from this and compared neural specificity metrics between connectome and equivalent random networks across tasks in memory, decision-making, and time-series prediction. Connectome ESNs contain smaller subsets of task-selective neurons, while random networks exhibit more distributed, “general” groups. We tested the interpretation of these metrics by systematically pruning nodes based on their measure of engagement, finding that connectome ESNs maintain performance more robustly across pruning. Finally, we investigated structural features of the networks, uncovering correlations between task-relevance and characteristics such as recurrence, node degree, and biological cell-type annotations. We find that correlations are consistent across connectome and conventional ESNs, but that connectome ESNs exhibit stronger correlations. These findings indicate that biologically-inspired connectivity can enable sparsely selective and compact neural networks, which may reduce energy consumption and optimise robustness. The approaches and metrics used in this research may suggest a way to initialise better performing, more efficient and robust ESNs, as well as provide insight into individual node feature importance in biological networks.

This talk is part of the Computational Neuroscience series.

Tell a friend about this talk:

This talk is included in these lists:

Note that ex-directory lists are not shown.

Other lists

Why Deep Neural Networks Are Promising for Speech Recognition Department of Archaeology LGBT+ History Month lecture The History of Science

Other talks

Poster Session Structural and molecular basis of DNA transfer by bacterial conjugative systems Recreating the Physical Natural World from Images The Prospects of International Cooperation on Phasing Out Fossil Fuels Gravitational Wave Signatures of Dark Matter in Neutron Star Mergers Gates Cambridge presents Dr Leor Zmigrod, author of 'The Ideological Brain'
 

© 2006-2025 Talks.cam, University of Cambridge. Contact Us | Help and Documentation | Privacy and Publicity