Modelling relational data using Bayesian clustered tensor factorization
We consider the problem of learning probabilistic models for complex relational structures between various types of objects. A model can help us "understand" a dataset of relational facts in at least two ways, by finding interpretable structure in the data, and by supporting predictions, o...
| Main Authors: | , , |
|---|---|
| Other Authors: | |
| Format: | Article |
| Published: |
Neural Information Processing Systems Foundation
2017
|
| Online Access: | http://hdl.handle.net/1721.1/112916 |
| _version_ | 1826212107433017344 |
|---|---|
| author | Sutskever, Ilya Tenenbaum, Joshua B. Salakhutdinov, Ruslan |
| author2 | Massachusetts Institute of Technology. Department of Brain and Cognitive Sciences |
| author_facet | Massachusetts Institute of Technology. Department of Brain and Cognitive Sciences Sutskever, Ilya Tenenbaum, Joshua B. Salakhutdinov, Ruslan |
| author_sort | Sutskever, Ilya |
| collection | MIT |
| description | We consider the problem of learning probabilistic models for complex relational structures between various types of objects. A model can help us "understand" a dataset of relational facts in at least two ways, by finding interpretable structure in the data, and by supporting predictions, or inferences about whether particular unobserved relations are likely to be true. Often there is a tradeoff between these two aims: cluster-based models yield more easily interpretable representations, while factorization-based approaches have given better predictive performance on large data sets. We introduce the Bayesian Clustered Tensor Factorization (BCTF) model, which embeds a factorized representation of relations in a nonparametric Bayesian clustering framework. Inference is fully Bayesian but scales well to large data sets. The model simultaneously discovers interpretable clusters and yields predictive performance that matches or beats previous probabilistic models for relational data. |
| first_indexed | 2024-09-23T15:16:30Z |
| format | Article |
| id | mit-1721.1/112916 |
| institution | Massachusetts Institute of Technology |
| last_indexed | 2024-09-23T15:16:30Z |
| publishDate | 2017 |
| publisher | Neural Information Processing Systems Foundation |
| record_format | dspace |
| spelling | mit-1721.1/1129162022-10-02T01:50:50Z Modelling relational data using Bayesian clustered tensor factorization Sutskever, Ilya Tenenbaum, Joshua B. Salakhutdinov, Ruslan Massachusetts Institute of Technology. Department of Brain and Cognitive Sciences Salakhutdinov, Ruslan We consider the problem of learning probabilistic models for complex relational structures between various types of objects. A model can help us "understand" a dataset of relational facts in at least two ways, by finding interpretable structure in the data, and by supporting predictions, or inferences about whether particular unobserved relations are likely to be true. Often there is a tradeoff between these two aims: cluster-based models yield more easily interpretable representations, while factorization-based approaches have given better predictive performance on large data sets. We introduce the Bayesian Clustered Tensor Factorization (BCTF) model, which embeds a factorized representation of relations in a nonparametric Bayesian clustering framework. Inference is fully Bayesian but scales well to large data sets. The model simultaneously discovers interpretable clusters and yields predictive performance that matches or beats previous probabilistic models for relational data. Natural Sciences and Engineering Research Council of Canada Shell Oil Company NTT Communication Science Laboratories United States. Air Force. Office of Scientific Research (FA9550-07-1-0075) United States. Air Force. Office of Scientific Research. Multidisciplinary University Research Initiative 2017-12-21T14:36:19Z 2017-12-21T14:36:19Z 2009-12 2017-12-08T18:37:32Z Article http://purl.org/eprint/type/ConferencePaper 1049-5258 http://hdl.handle.net/1721.1/112916 Sutskever, Ilya, Ruslan Salakhutdinov and Joshua B. Tenenbaum. "Modelling Relational Data using Bayesian Clustered Tensor Factorization." Advances in Neural Information Processing Systems 22 (NIPS 2009), Vancouver, British Columbia, Canada, 6-10 December, 2009. © 2009 Neural Information Processing Systems Foundation https://papers.nips.cc/paper/3863-modelling-relational-data-using-bayesian-clustered-tensor-factorization Advances in Neural Information Processing Systems 22 (NIPS 2009) Article is made available in accordance with the publisher's policy and may be subject to US copyright law. Please refer to the publisher's site for terms of use. application/pdf Neural Information Processing Systems Foundation Neural Information Processing Systems (NIPS) |
| spellingShingle | Sutskever, Ilya Tenenbaum, Joshua B. Salakhutdinov, Ruslan Modelling relational data using Bayesian clustered tensor factorization |
| title | Modelling relational data using Bayesian clustered tensor factorization |
| title_full | Modelling relational data using Bayesian clustered tensor factorization |
| title_fullStr | Modelling relational data using Bayesian clustered tensor factorization |
| title_full_unstemmed | Modelling relational data using Bayesian clustered tensor factorization |
| title_short | Modelling relational data using Bayesian clustered tensor factorization |
| title_sort | modelling relational data using bayesian clustered tensor factorization |
| url | http://hdl.handle.net/1721.1/112916 |
| work_keys_str_mv | AT sutskeverilya modellingrelationaldatausingbayesianclusteredtensorfactorization AT tenenbaumjoshuab modellingrelationaldatausingbayesianclusteredtensorfactorization AT salakhutdinovruslan modellingrelationaldatausingbayesianclusteredtensorfactorization |