In this paper we demonstrate how deterministic annealing can be applied to different SVM formulations of the multiple-instance learning (MIL) problem. Our results show that we find better local minima compared to the heuristic methods those problems are usually solved with. However this does not always translate into a better test error suggesting an inadequacy of the objective function. Based on this finding we propose a new objective function which together with the deterministic annealing algorithm finds better local minima and achieves better performance on a set of benchmark datasets. Furthermore the results also show how the structure of MIL datasets influence the performance of MIL algorithms and we discuss how future benchmark datasets for the MIL problem should be designed.
| Author(s): | Gehler, PV. and Chapelle, O. |
| Book Title: | JMLR Workshop and Conference Proceedings Volume 2: AISTATS 2007 |
| Journal: | Proceedings of the 11th International Conference on Artificial Intelligence and Statistics (AISTATS 2007) |
| Pages: | 123-130 |
| Year: | 2007 |
| Month: | March |
| Day: | 0 |
| Editors: | Meila, M. , X. Shen |
| Publisher: | MIT Press |
| Bibtex Type: | Conference Paper (inproceedings) |
| Address: | Cambridge, MA, USA |
| Event Name: | 11th International Conference on Artificial Intelligence and Statistics |
| Event Place: | San Juan, Puerto Rico |
| Digital: | 0 |
| Electronic Archiving: | grant_archive |
| Language: | en |
| Organization: | Max-Planck-Gesellschaft |
| School: | Biologische Kybernetik |
| Links: | |
BibTex
@inproceedings{4270,
title = {Deterministic Annealing for
Multiple-Instance Learning},
journal = {Proceedings of the 11th International Conference on Artificial Intelligence and Statistics (AISTATS 2007)},
booktitle = {JMLR Workshop and Conference Proceedings Volume 2: AISTATS 2007},
abstract = {In this paper we demonstrate how deterministic annealing can be applied to different SVM formulations of the multiple-instance learning (MIL) problem. Our results show that we find better local minima compared to the heuristic methods those problems are usually solved with. However this does not always translate into a better test error suggesting an inadequacy of the objective function. Based on this finding we propose a new objective function which together with the deterministic annealing algorithm finds better local minima and achieves better performance on a set of benchmark datasets. Furthermore the results also show how the structure of MIL datasets influence the performance of MIL algorithms and we discuss how future benchmark datasets for the MIL problem should be designed. },
pages = {123-130},
editors = {Meila, M. , X. Shen},
publisher = {MIT Press},
organization = {Max-Planck-Gesellschaft},
school = {Biologische Kybernetik},
address = {Cambridge, MA, USA},
month = mar,
year = {2007},
slug = {4270},
author = {Gehler, PV. and Chapelle, O.},
month_numeric = {3}
}