On Effects of Compression with Hyperdimensional Computing in Distributed Randomized Neural Networks
2021 (English)In: Advances in Computational Intelligence: 16th International Work-Conference on Artificial Neural Networks, IWANN 2021, Virtual Event, June 16–18, 2021, Proceedings, Part II / [ed] Ignacio Rojas; Gonzalo Joya; Andreu Catala, Springer, 2021, p. 155-167Conference paper, Published paper (Refereed)
Abstract [en]
A change of the prevalent supervised learning techniques is foreseeable in the near future: from the complex, computational expensive algorithms to more flexible and elementary training ones. The strong revitalization of randomized algorithms can be framed in this prospect steering. We recently proposed a model for distributed classification based on randomized neural networks and hyperdimensional computing, which takes into account cost of information exchange between agents using compression. The use of compression is important as it addresses the issues related to the communication bottleneck, however, the original approach is rigid in the way the compression is used. Therefore, in this work, we propose a more flexible approach to compression and compare it to conventional compression algorithms, dimensionality reduction, and quantization techniques.
Place, publisher, year, edition, pages
Springer, 2021. p. 155-167
Series
Lecture Notes in Computer Science, ISSN 0302-9743, E-ISSN 1611-3349 ; 12862
Keywords [en]
Distributed randomized neural networks, Compression, Vector symbolic architectures, Hyperdimensional computing
National Category
Computer Sciences
Research subject
Dependable Communication and Computation Systems
Identifiers
URN: urn:nbn:se:ltu:diva-87253DOI: 10.1007/978-3-030-85099-9_13ISI: 000696688800013Scopus ID: 2-s2.0-85115177088OAI: oai:DiVA.org:ltu-87253DiVA, id: diva2:1597920
Conference
16th International Work-Conference on Artificial Neural Networks (IWANN 2021), online, June 16-18, 2021
Funder
EU, Horizon 2020, 839179
Note
ISBN för värdpublikation: 978-3-030-85098-2; 978-3-030-85099-9;
Funder: DARPA’s AIE (HyDDENN) program; AFOSR (FA9550-19-1-0241)
2021-09-282021-09-282021-10-08Bibliographically approved