Dynamic Graph Structure Estimation for Learning Multivariate Point Process using Spiking Neural Networks

Biswadeep Chakraborty, Hemant Kumawat, Beomseok Kang, Saibal Mukhopadhyay · 2025

Read Dynamic Graph Structure Estimation for Learning Multivariate Point Process using Spiking Neural Networks by Biswadeep Chakraborty, Hemant Kumawat, Beomseok Kang, Saibal Mukhopadhyay (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2504-01246.txt.

Description

Biswadeep Chakraborty and co-authors present a spiking network for learning event data in a 2025 paper. Many fields, from brain science to finance, need models of events that fire at uneven times across many linked streams. The links between streams shift over time, yet most models fix the link map in advance or take it as given.

Their network uses spiking units, which fire in pulses like brain cells, and a learning rule that rewards near-simultaneous firing. From the event data alone it builds a map of links that can change as new events arrive. Tests on made-up data and on real sets, from taxi trips to forum posts, gave strong forecasts with modest computing cost. Checks that removed parts of the design showed each part helped.

Catalog record

CREATORBiswadeep Chakraborty, Hemant Kumawat, Beomseok Kang, Saibal Mukhopadhyay
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE81 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2504-01246

Browse the full library · About the Tower of Babel