Adaptive-Size Dictionary Learning Using Information Theoretic Criteria
Finding the size of the dictionary is an open issue in dictionary learning (DL). We propose an algorithm that adapts the size during the learning process by using Information Theoretic Criteria (ITC) specialized to the DL problem. The algorithm is built on top of Approximate K-SVD (AK-SVD) and perio...
| Main Authors: | , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2019-08-01
|
| Series: | Algorithms |
| Subjects: | |
| Online Access: | https://www.mdpi.com/1999-4893/12/9/178 |
| _version_ | 1818191171108732928 |
|---|---|
| author | Bogdan Dumitrescu Ciprian Doru Giurcăneanu |
| author_facet | Bogdan Dumitrescu Ciprian Doru Giurcăneanu |
| author_sort | Bogdan Dumitrescu |
| collection | DOAJ |
| description | Finding the size of the dictionary is an open issue in dictionary learning (DL). We propose an algorithm that adapts the size during the learning process by using Information Theoretic Criteria (ITC) specialized to the DL problem. The algorithm is built on top of Approximate K-SVD (AK-SVD) and periodically removes the less used atoms or adds new random atoms, based on ITC evaluations for a small number of candidate sub-dictionaries. Numerical experiments on synthetic data show that our algorithm not only finds the true size with very good accuracy, but is also able to improve the representation error in comparison with AK-SVD knowing the true size. |
| first_indexed | 2024-12-12T00:10:22Z |
| format | Article |
| id | doaj.art-45eb107936db4ce9b5aa597a6634eaac |
| institution | Directory Open Access Journal |
| issn | 1999-4893 |
| language | English |
| last_indexed | 2024-12-12T00:10:22Z |
| publishDate | 2019-08-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Algorithms |
| spelling | doaj.art-45eb107936db4ce9b5aa597a6634eaac2022-12-22T00:45:00ZengMDPI AGAlgorithms1999-48932019-08-0112917810.3390/a12090178a12090178Adaptive-Size Dictionary Learning Using Information Theoretic CriteriaBogdan Dumitrescu0Ciprian Doru Giurcăneanu1Department of Automatic Control and Computers, University Politehnica of Bucharest, 313 Spl. Independenţei, 060042 Bucharest, RomaniaDepartment of Statistics, University of Auckland, Auckland 1142, New ZealandFinding the size of the dictionary is an open issue in dictionary learning (DL). We propose an algorithm that adapts the size during the learning process by using Information Theoretic Criteria (ITC) specialized to the DL problem. The algorithm is built on top of Approximate K-SVD (AK-SVD) and periodically removes the less used atoms or adds new random atoms, based on ITC evaluations for a small number of candidate sub-dictionaries. Numerical experiments on synthetic data show that our algorithm not only finds the true size with very good accuracy, but is also able to improve the representation error in comparison with AK-SVD knowing the true size.https://www.mdpi.com/1999-4893/12/9/178dictionary learningsparse representationinformation theoretic criteriadictionary size |
| spellingShingle | Bogdan Dumitrescu Ciprian Doru Giurcăneanu Adaptive-Size Dictionary Learning Using Information Theoretic Criteria Algorithms dictionary learning sparse representation information theoretic criteria dictionary size |
| title | Adaptive-Size Dictionary Learning Using Information Theoretic Criteria |
| title_full | Adaptive-Size Dictionary Learning Using Information Theoretic Criteria |
| title_fullStr | Adaptive-Size Dictionary Learning Using Information Theoretic Criteria |
| title_full_unstemmed | Adaptive-Size Dictionary Learning Using Information Theoretic Criteria |
| title_short | Adaptive-Size Dictionary Learning Using Information Theoretic Criteria |
| title_sort | adaptive size dictionary learning using information theoretic criteria |
| topic | dictionary learning sparse representation information theoretic criteria dictionary size |
| url | https://www.mdpi.com/1999-4893/12/9/178 |
| work_keys_str_mv | AT bogdandumitrescu adaptivesizedictionarylearningusinginformationtheoreticcriteria AT cipriandorugiurcaneanu adaptivesizedictionarylearningusinginformationtheoreticcriteria |