Vis enkel innførsel

dc.contributor.authorAbeyrathna, Kuruge Darshana
dc.contributor.authorBhattarai, Bimal
dc.contributor.authorGoodwin, Morten
dc.contributor.authorGorji, Saeed Rahimi
dc.contributor.authorGranmo, Ole-Christoffer
dc.contributor.authorLei, Jiao
dc.contributor.authorSaha, Rupsa
dc.contributor.authorYadav, Rohan Kumar
dc.date.accessioned2022-04-28T09:28:47Z
dc.date.available2022-04-28T09:28:47Z
dc.date.created2022-01-06T16:18:45Z
dc.date.issued2021
dc.identifier.citationAbeyrathna, K. D., Bhattarai, B., Goodwin, M., Gorji, S. R., Granmo, O.-C., Lei, J., Saha, R. & Yadav, R. K. (2021). Massively Parallel and Asynchronous Tsetlin Machine Architecture Supporting Almost Constant-Time Scaling. Proceedings of Machine Learning Research (PMLR), 11.en_US
dc.identifier.issn2640-3498
dc.identifier.urihttps://hdl.handle.net/11250/2993140
dc.description.abstractUsing logical clauses to represent patterns, Tsetlin Machine (TM) have recently obtained competitive performance in terms of accuracy, memory footprint, energy, and learning speed on several benchmarks. Each TM clause votes for or against a particular class, with classification resolved using a majority vote. While the evaluation of clauses is fast, being based on binary operators, the voting makes it necessary to synchronize the clause evaluation, impeding parallelization. In this paper, we propose a novel scheme for desynchronizing the evaluation of clauses, eliminating the voting bottleneck. In brief, every clause runs in its own thread for massive native parallelism. For each training example, we keep track of the class votes obtained from the clauses in local voting tallies. The local voting tallies allow us to detach the processing of each clause from the rest of the clauses, supporting decentralized learning. This means that the TM most of the time will operate on outdated voting tallies. We evaluated the proposed parallelization across diverse learning tasks and it turns out that our decentralized TM learning algorithm copes well with working on outdated data, resulting in no significant loss in learning accuracy. Furthermore, we show that the approach provides up to 50 times faster learning. Finally, learning time is almost constant for reasonable clause amounts (employing from 20 to 7,000 clauses on a Tesla V100 GPU). For sufficiently large clause numbers, computation time increases approximately proportionally. Our parallel and asynchronous architecture thus allows processing of more massive datasets and operating with more clauses for higher accuracy.en_US
dc.language.isoengen_US
dc.publisherJMLRen_US
dc.relation.urihttp://proceedings.mlr.press/v139/abeyrathna21a/abeyrathna21a.pdf
dc.titleMassively Parallel and Asynchronous Tsetlin Machine Architecture Supporting Almost Constant-Time Scalingen_US
dc.typePeer revieweden_US
dc.typeJournal articleen_US
dc.description.versionsubmittedVersionen_US
dc.rights.holder© 2021 The Authorsen_US
dc.subject.nsiVDP::Teknologi: 500::Informasjons- og kommunikasjonsteknologi: 550en_US
dc.source.pagenumber11en_US
dc.source.journalProceedings of Machine Learning Research (PMLR)en_US
dc.identifier.cristin1976140
dc.relation.projectUniversitetet i Agder: CAIRen_US
cristin.qualitycode1


Tilhørende fil(er)

Thumbnail

Denne innførselen finnes i følgende samling(er)

Vis enkel innførsel