{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,5,3]],"date-time":"2026-05-03T03:18:56Z","timestamp":1777778336485,"version":"3.51.4"},"reference-count":34,"publisher":"SAGE Publications","issue":"1","license":[{"start":{"date-parts":[[2022,10,27]],"date-time":"2022-10-27T00:00:00Z","timestamp":1666828800000},"content-version":"tdm","delay-in-days":0,"URL":"https:\/\/journals.sagepub.com\/page\/policies\/text-and-data-mining-license"}],"content-domain":{"domain":["journals.sagepub.com"],"crossmark-restriction":true},"short-container-title":["Information Visualization"],"published-print":{"date-parts":[[2023,1]]},"abstract":"<jats:p>Word embedding, a high-dimensional (HD) numerical representation of words generated by machine learning models, has been used for different natural language processing tasks, for example, translation between two languages. Recently, there has been an increasing trend of transforming the HD embeddings into a latent space (e.g. via autoencoders) for further tasks, exploiting various merits the latent representations could bring. To preserve the embeddings\u2019 quality, these works often map the embeddings into an even higher-dimensional latent space, making the already complicated embeddings even less interpretable and consuming more storage space. In this work, we borrow the idea of [Formula: see text] VAE to regularize the HD latent space. Our regularization implicitly condenses information from the HD latent space into a much lower-dimensional space, thus compressing the embeddings. We also show that each dimension of our regularized latent space is more semantically salient, and validate our assertion by interactively probing the encoding-level of user-proposed semantics in the dimensions. To the end, we design a visual analytics system to monitor the regularization process, explore the HD latent space, and interpret latent dimensions\u2019 semantics. We validate the effectiveness of our embedding regularization and interpretation approach through both quantitative and qualitative evaluations.<\/jats:p>","DOI":"10.1177\/14738716221130338","type":"journal-article","created":{"date-parts":[[2022,10,27]],"date-time":"2022-10-27T07:01:47Z","timestamp":1666854107000},"page":"52-68","update-policy":"https:\/\/doi.org\/10.1177\/sage-journals-update-policy","source":"Crossref","is-referenced-by-count":3,"title":["Compressing and interpreting word embeddings with latent space regularization and interactive semantics probing"],"prefix":"10.1177","volume":"22","author":[{"ORCID":"https:\/\/orcid.org\/0000-0002-7138-8263","authenticated-orcid":false,"given":"Haoyu","family":"Li","sequence":"first","affiliation":[{"name":"GRAVITY Research Group, Department of Computer Science and Engineering, The Ohio State University, Columbus, OH, USA"},{"name":"Visa Research, Palo Alto, CA, USA"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Junpeng","family":"Wang","sequence":"additional","affiliation":[{"name":"Visa Research, Palo Alto, CA, USA"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Yan","family":"Zheng","sequence":"additional","affiliation":[{"name":"Visa Research, Palo Alto, CA, USA"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Liang","family":"Wang","sequence":"additional","affiliation":[{"name":"Visa Research, Palo Alto, CA, USA"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Wei","family":"Zhang","sequence":"additional","affiliation":[{"name":"Visa Research, Palo Alto, CA, USA"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Han-Wei","family":"Shen","sequence":"additional","affiliation":[{"name":"GRAVITY Research Group, Department of Computer Science and Engineering, The Ohio State University, Columbus, OH, USA"}],"role":[{"role":"author","vocabulary":"crossref"}]}],"member":"179","published-online":{"date-parts":[[2022,10,27]]},"reference":[{"key":"bibr1-14738716221130338","doi-asserted-by":"publisher","DOI":"10.1109\/TVCG.2017.2745141"},{"key":"bibr2-14738716221130338","first-page":"48","volume-title":"2018 IEEE VAST","author":"Li Q"},{"key":"bibr3-14738716221130338","doi-asserted-by":"publisher","DOI":"10.1109\/TVCG.2019.2903946"},{"key":"bibr4-14738716221130338","author":"Mohiuddin T","year":"2020","journal-title":"arXiv preprint"},{"key":"bibr5-14738716221130338","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-319-57959-7"},{"key":"bibr6-14738716221130338","author":"Kingma DP","year":"2013","journal-title":"arXiv preprint"},{"key":"bibr7-14738716221130338","volume-title":"5th International conference on learning representations","author":"Higgins I","year":"2017"},{"key":"bibr8-14738716221130338","author":"Higgins I","year":"2017","journal-title":"arXiv preprint"},{"key":"bibr9-14738716221130338","first-page":"51","volume-title":"IEEE Pacific visualization symposium","author":"Wang J"},{"key":"bibr10-14738716221130338","first-page":"95","author":"Heinrich J","journal-title":"Eurographics (State of the Art Reports)"},{"key":"bibr11-14738716221130338","doi-asserted-by":"publisher","DOI":"10.1109\/TVCG.2008.153"},{"key":"bibr12-14738716221130338","first-page":"437","volume-title":"Proceedings. Visualization\u201997 (Cat. No. 97CB36155)","author":"Hoffman P"},{"key":"bibr13-14738716221130338","first-page":"22","volume-title":"Proceedings of the IEEE information visualization symposium","volume":"650","author":"Kandogan E"},{"issue":"11","key":"bibr14-14738716221130338","first-page":"2579","volume":"9","author":"Van der Maaten L","year":"2008","journal-title":"J Mach Learn Res"},{"key":"bibr15-14738716221130338","author":"McInnes L","year":"2018","journal-title":"arXiv preprint"},{"key":"bibr16-14738716221130338","doi-asserted-by":"publisher","DOI":"10.1109\/MCG.2018.042731661"},{"key":"bibr17-14738716221130338","doi-asserted-by":"publisher","DOI":"10.1007\/s41095-020-0191-7"},{"key":"bibr18-14738716221130338","first-page":"12","volume-title":"2020 IEEE conference on visual analytics science and technology (VAST)","author":"Yang W"},{"key":"bibr19-14738716221130338","doi-asserted-by":"publisher","DOI":"10.1109\/TVCG.2016.2598831"},{"key":"bibr20-14738716221130338","doi-asserted-by":"publisher","DOI":"10.1109\/TVCG.2016.2598828"},{"key":"bibr21-14738716221130338","author":"Rathore A","year":"2021","journal-title":"arXiv preprint"},{"issue":"1","key":"bibr22-14738716221130338","first-page":"1001","volume":"26","author":"El-Assady M","year":"2020","journal-title":"IEEE Trans Vis Comput Graph"},{"key":"bibr23-14738716221130338","doi-asserted-by":"publisher","DOI":"10.1109\/TVCG.2017.2744478"},{"key":"bibr24-14738716221130338","doi-asserted-by":"publisher","DOI":"10.1111\/cgf.13672"},{"key":"bibr25-14738716221130338","doi-asserted-by":"publisher","DOI":"10.1109\/TVCG.2019.2903943"},{"key":"bibr26-14738716221130338","doi-asserted-by":"publisher","DOI":"10.1109\/TVCG.2020.3030350"},{"key":"bibr27-14738716221130338","author":"Burgess CP","year":"2018","journal-title":"arXiv preprint"},{"key":"bibr28-14738716221130338","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/S17-2002"},{"key":"bibr29-14738716221130338","author":"Mikolov T","year":"2013","journal-title":"arXiv preprint"},{"key":"bibr30-14738716221130338","author":"Conneau A","year":"2017","journal-title":"arXiv preprint"},{"key":"bibr31-14738716221130338","author":"Mikolov T","year":"2013","journal-title":"arXiv preprint"},{"key":"bibr32-14738716221130338","doi-asserted-by":"publisher","DOI":"10.1613\/jair.1.11640"},{"key":"bibr33-14738716221130338","doi-asserted-by":"publisher","DOI":"10.1162\/tacl_a_00051"},{"key":"bibr34-14738716221130338","first-page":"361","volume-title":"Proceedings of the First IEEE conference on visualization","author":"Inselberg A"}],"container-title":["Information Visualization"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/journals.sagepub.com\/doi\/pdf\/10.1177\/14738716221130338","content-type":"application\/pdf","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/journals.sagepub.com\/doi\/full-xml\/10.1177\/14738716221130338","content-type":"application\/xml","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/journals.sagepub.com\/doi\/pdf\/10.1177\/14738716221130338","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,4,29]],"date-time":"2026-04-29T19:19:15Z","timestamp":1777490355000},"score":1,"resource":{"primary":{"URL":"https:\/\/journals.sagepub.com\/doi\/10.1177\/14738716221130338"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2022,10,27]]},"references-count":34,"journal-issue":{"issue":"1","published-print":{"date-parts":[[2023,1]]}},"alternative-id":["10.1177\/14738716221130338"],"URL":"https:\/\/doi.org\/10.1177\/14738716221130338","relation":{},"ISSN":["1473-8716","1473-8724"],"issn-type":[{"value":"1473-8716","type":"print"},{"value":"1473-8724","type":"electronic"}],"subject":[],"published":{"date-parts":[[2022,10,27]]}}}