A Statistical Learning Theory Framework for Supervised Pattern Discovery
Copyright © SIAM. This paper formalizes a latent variable inference problem we call supervised, pattern discovery, the goal of which is to find sets of observations that belong to a single "pattern." We discuss two versions of the problem and prove uniform risk bounds for both. In the firs...
| Main Authors: | , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Society for Industrial and Applied Mathematics
2021
|
| Online Access: | https://hdl.handle.net/1721.1/137446 |
| _version_ | 1826216302551760896 |
|---|---|
| author | Huggins, Jonathan H. Rudin, Cynthia |
| author_facet | Huggins, Jonathan H. Rudin, Cynthia |
| author_sort | Huggins, Jonathan H. |
| collection | MIT |
| description | Copyright © SIAM. This paper formalizes a latent variable inference problem we call supervised, pattern discovery, the goal of which is to find sets of observations that belong to a single "pattern." We discuss two versions of the problem and prove uniform risk bounds for both. In the first version, collections of patterns can be generated in an arbitrary manner and the data consist of multiple labeled collections. In the second version, the patterns are assumed to be generated independently by identically distributed processes. These processes are allowed to take an arbitrary form, so observations within a pattern are not in general independent of each other. The bounds for the second version of the problem are stated in terms of a new complexity measure, the quasi-Rademacher complexity. |
| first_indexed | 2024-09-23T16:45:15Z |
| format | Article |
| id | mit-1721.1/137446 |
| institution | Massachusetts Institute of Technology |
| language | English |
| last_indexed | 2024-09-23T16:45:15Z |
| publishDate | 2021 |
| publisher | Society for Industrial and Applied Mathematics |
| record_format | dspace |
| spelling | mit-1721.1/1374462021-11-06T03:01:28Z A Statistical Learning Theory Framework for Supervised Pattern Discovery Huggins, Jonathan H. Rudin, Cynthia Copyright © SIAM. This paper formalizes a latent variable inference problem we call supervised, pattern discovery, the goal of which is to find sets of observations that belong to a single "pattern." We discuss two versions of the problem and prove uniform risk bounds for both. In the first version, collections of patterns can be generated in an arbitrary manner and the data consist of multiple labeled collections. In the second version, the patterns are assumed to be generated independently by identically distributed processes. These processes are allowed to take an arbitrary form, so observations within a pattern are not in general independent of each other. The bounds for the second version of the problem are stated in terms of a new complexity measure, the quasi-Rademacher complexity. 2021-11-05T13:16:30Z 2021-11-05T13:16:30Z 2014-04-28 2019-05-10T12:29:19Z Article http://purl.org/eprint/type/ConferencePaper https://hdl.handle.net/1721.1/137446 Huggins, Jonathan H. and Rudin, Cynthia. 2014. "A Statistical Learning Theory Framework for Supervised Pattern Discovery." en 10.1137/1.9781611973440.58 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 Society for Industrial and Applied Mathematics SIAM |
| spellingShingle | Huggins, Jonathan H. Rudin, Cynthia A Statistical Learning Theory Framework for Supervised Pattern Discovery |
| title | A Statistical Learning Theory Framework for Supervised Pattern Discovery |
| title_full | A Statistical Learning Theory Framework for Supervised Pattern Discovery |
| title_fullStr | A Statistical Learning Theory Framework for Supervised Pattern Discovery |
| title_full_unstemmed | A Statistical Learning Theory Framework for Supervised Pattern Discovery |
| title_short | A Statistical Learning Theory Framework for Supervised Pattern Discovery |
| title_sort | statistical learning theory framework for supervised pattern discovery |
| url | https://hdl.handle.net/1721.1/137446 |
| work_keys_str_mv | AT hugginsjonathanh astatisticallearningtheoryframeworkforsupervisedpatterndiscovery AT rudincynthia astatisticallearningtheoryframeworkforsupervisedpatterndiscovery AT hugginsjonathanh statisticallearningtheoryframeworkforsupervisedpatterndiscovery AT rudincynthia statisticallearningtheoryframeworkforsupervisedpatterndiscovery |