Abstract
We present and analyze a theoretical model designed to understand and explain the effectiveness of ontologies for learning multiple related tasks from primarily unlabeled data. We present both information-theoretic results as well as efficient algorithms. We show in this model that an ontology, which specifies the relationships between multiple outputs, in some cases is sufficient to completely learn a classification using a large unlabeled data source.
| Original language | English |
|---|---|
| Pages | 2149-2157 |
| Number of pages | 9 |
| State | Published - 2013 |
| Event | 30th International Conference on Machine Learning, ICML 2013 - Atlanta, GA, United States Duration: 16 Jun 2013 → 21 Jun 2013 |
Conference
| Conference | 30th International Conference on Machine Learning, ICML 2013 |
|---|---|
| Country/Territory | United States |
| City | Atlanta, GA |
| Period | 16/06/13 → 21/06/13 |
Funding
| Funders | Funder number |
|---|---|
| Air Force Office of Scientific Research | FA9550-09-1-0538 |
| National Science Foundation | CCF-0953192, CCF-1116892, IIS-1065251 |
| Office of Naval Research | N00014-09-1-0751 |