DC Field | Value | Language |
---|---|---|
dc.contributor.author | Zhang, Jiyuan | ko |
dc.contributor.author | Oh, Jinoh | ko |
dc.contributor.author | Shin, Kijung | ko |
dc.contributor.author | Papalexakis, Evangelos E. | ko |
dc.contributor.author | Faloutsos, Christos | ko |
dc.contributor.author | Yu, Hwanjo | ko |
dc.date.accessioned | 2020-10-20T01:55:32Z | - |
dc.date.available | 2020-10-20T01:55:32Z | - |
dc.date.created | 2020-01-23 | - |
dc.date.created | 2020-01-23 | - |
dc.date.created | 2020-01-23 | - |
dc.date.created | 2020-01-23 | - |
dc.date.issued | 2020-07 | - |
dc.identifier.citation | KNOWLEDGE AND INFORMATION SYSTEMS, v.62, no.7, pp.2765 - 2794 | - |
dc.identifier.issn | 0219-1377 | - |
dc.identifier.uri | http://hdl.handle.net/10203/276684 | - |
dc.description.abstract | Multi-aspect data appear frequently in web-related applications. For example, product reviews are quadruplets of the form (user, product, keyword, timestamp), and search-engine logs are quadruplets of the form (user, keyword, location, timestamp). How can we analyze such web-scale multi-aspect data on an off-the-shelf workstation with a limited amount of memory? Tucker decomposition has been used widely for discovering patterns in such multi-aspect data, which are naturally expressed as large but sparse tensors. However, existing Tucker decomposition algorithms have limited scalability, failing to decompose large-scale high-order (= 4) tensors, since they explicitly materialize intermediate data, whose size grows exponentially with the order. To address this problem, which we call "Materialization Bottleneck," we propose S- HOT, a scalable algorithm for high-order Tucker decomposition. S- HOT minimizes materialized intermediate data by using an on-the-fly computation, and it is optimized for disk-resident tensors that are too large to fit in memory. We theoretically analyze the amount of memory and the number of data scans required by S- HOT. Moreover, we empirically showthat S- HOT handles tensors with higher order, dimensionality, and rank than baselines. For example, S- HOT successfully decomposes a real-world tensor from the Microsoft Academic Graph on an off-the-shelf workstation, while all baselines fail. Especially, in terms of dimensionality, S- HOT decomposes 1000xlarger tensors than baselines. | - |
dc.language | English | - |
dc.publisher | SPRINGER LONDON LTD | - |
dc.title | Fast and Memory-Efficient Algorithms for High-Order Tucker Decomposition | - |
dc.type | Article | - |
dc.identifier.wosid | 000515603900001 | - |
dc.identifier.scopusid | 2-s2.0-85078491286 | - |
dc.type.rims | ART | - |
dc.citation.volume | 62 | - |
dc.citation.issue | 7 | - |
dc.citation.beginningpage | 2765 | - |
dc.citation.endingpage | 2794 | - |
dc.citation.publicationname | KNOWLEDGE AND INFORMATION SYSTEMS | - |
dc.identifier.doi | 10.1007/s10115-019-01435-1 | - |
dc.contributor.localauthor | Shin, Kijung | - |
dc.contributor.nonIdAuthor | Zhang, Jiyuan | - |
dc.contributor.nonIdAuthor | Oh, Jinoh | - |
dc.contributor.nonIdAuthor | Papalexakis, Evangelos E. | - |
dc.contributor.nonIdAuthor | Faloutsos, Christos | - |
dc.contributor.nonIdAuthor | Yu, Hwanjo | - |
dc.description.isOpenAccess | N | - |
dc.type.journalArticle | Article | - |
dc.subject.keywordAuthor | Tensor | - |
dc.subject.keywordAuthor | High-order tensor | - |
dc.subject.keywordAuthor | Tensor decomposition | - |
dc.subject.keywordAuthor | Tucker decomposition | - |
dc.subject.keywordAuthor | Out-of-core algorithm | - |
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.