Generalization bounds for learning with linear, polygonal, quadratic and conic side knowledge
In this paper, we consider a supervised learning setting where side knowledge is provided about the labels of unlabeled examples. The side knowledge has the effect of reducing the hypothesis space, leading to tighter generalization bounds, and thus possibly better generalization. We consider several...
| Main Authors: | , |
|---|---|
| Other Authors: | |
| Format: | Article |
| Language: | English |
| Published: |
Springer Science+Business Media
2016
|
| Online Access: | http://hdl.handle.net/1721.1/103110 |
| _version_ | 1811096766393614336 |
|---|---|
| author | Tulabandhula, Theja Rudin, Cynthia |
| author2 | Massachusetts Institute of Technology. Department of Electrical Engineering and Computer Science |
| author_facet | Massachusetts Institute of Technology. Department of Electrical Engineering and Computer Science Tulabandhula, Theja Rudin, Cynthia |
| author_sort | Tulabandhula, Theja |
| collection | MIT |
| description | In this paper, we consider a supervised learning setting where side knowledge is provided about the labels of unlabeled examples. The side knowledge has the effect of reducing the hypothesis space, leading to tighter generalization bounds, and thus possibly better generalization. We consider several types of side knowledge, the first leading to linear and polygonal constraints on the hypothesis space, the second leading to quadratic constraints, and the last leading to conic constraints. We show how different types of domain knowledge can lead directly to these kinds of side knowledge. We prove bounds on complexity measures of the hypothesis space for quadratic and conic side knowledge, and show that these bounds are tight in a specific sense for the quadratic case. |
| first_indexed | 2024-09-23T16:48:42Z |
| format | Article |
| id | mit-1721.1/103110 |
| institution | Massachusetts Institute of Technology |
| language | English |
| last_indexed | 2024-09-23T16:48:42Z |
| publishDate | 2016 |
| publisher | Springer Science+Business Media |
| record_format | dspace |
| spelling | mit-1721.1/1031102022-09-29T21:39:07Z Generalization bounds for learning with linear, polygonal, quadratic and conic side knowledge Tulabandhula, Theja Rudin, Cynthia Massachusetts Institute of Technology. Department of Electrical Engineering and Computer Science Sloan School of Management Tulabandhula, Theja Rudin, Cynthia In this paper, we consider a supervised learning setting where side knowledge is provided about the labels of unlabeled examples. The side knowledge has the effect of reducing the hypothesis space, leading to tighter generalization bounds, and thus possibly better generalization. We consider several types of side knowledge, the first leading to linear and polygonal constraints on the hypothesis space, the second leading to quadratic constraints, and the last leading to conic constraints. We show how different types of domain knowledge can lead directly to these kinds of side knowledge. We prove bounds on complexity measures of the hypothesis space for quadratic and conic side knowledge, and show that these bounds are tight in a specific sense for the quadratic case. 2016-06-14T19:14:45Z 2016-06-14T19:14:45Z 2014-12 2013-12 2016-05-23T12:15:05Z Article http://purl.org/eprint/type/JournalArticle 0885-6125 1573-0565 http://hdl.handle.net/1721.1/103110 Tulabandhula, Theja, and Cynthia Rudin. "Generalization bounds for learning with linear, polygonal, quadratic and conic side knowledge." Machine Learning 100:2-3 (2015), pp.183-216. en http://dx.doi.org/10.1007/s10994-014-5478-4 Machine Learning Creative Commons Attribution-Noncommercial-Share Alike http://creativecommons.org/licenses/by-nc-sa/4.0/ The Author(s) application/pdf Springer Science+Business Media Springer US |
| spellingShingle | Tulabandhula, Theja Rudin, Cynthia Generalization bounds for learning with linear, polygonal, quadratic and conic side knowledge |
| title | Generalization bounds for learning with linear, polygonal, quadratic and conic side knowledge |
| title_full | Generalization bounds for learning with linear, polygonal, quadratic and conic side knowledge |
| title_fullStr | Generalization bounds for learning with linear, polygonal, quadratic and conic side knowledge |
| title_full_unstemmed | Generalization bounds for learning with linear, polygonal, quadratic and conic side knowledge |
| title_short | Generalization bounds for learning with linear, polygonal, quadratic and conic side knowledge |
| title_sort | generalization bounds for learning with linear polygonal quadratic and conic side knowledge |
| url | http://hdl.handle.net/1721.1/103110 |
| work_keys_str_mv | AT tulabandhulatheja generalizationboundsforlearningwithlinearpolygonalquadraticandconicsideknowledge AT rudincynthia generalizationboundsforlearningwithlinearpolygonalquadraticandconicsideknowledge |