Projekte pro Jahr
Abstract
The combination of two major challenges in algorithmic learning is investigated: dealing with huge amounts of irrelevant information and learning from noisy data. It is shown that large classes of Boolean concepts that only depend on a small fraction of their variables - so-called juntas - can be learned efficiently from uniformly distributed examples that are corrupted by random attribute and classification noise. We present solutions to cope with the manifold problems that inhibit a straightforward generalization of the noise-free case. Additionally, we extend our methods to non-uniformly distributed examples and derive new results for monotone juntas in this setting. We assume that the attribute noise is generated by a product distribution. Otherwise fault-tolerant learning is in general impossible which follows from the construction of a noise distribution P and a concept class C such that it is impossible to learn C under P-noise.
Originalsprache | Englisch |
---|---|
Titel | TAMC 2006: Theory and Applications of Models of Computation |
Seitenumfang | 12 |
Band | 3959 LNCS |
Herausgeber (Verlag) | Springer Berlin Heidelberg |
Erscheinungsdatum | 17.07.2006 |
Seiten | 387-398 |
ISBN (Print) | 978-3-540-34021-8 |
ISBN (elektronisch) | 978-3-540-34022-5 |
DOIs | |
Publikationsstatus | Veröffentlicht - 17.07.2006 |
Veranstaltung | 3rd International Conference on Theory and Applications of Models of Computation - Beijing, China Dauer: 15.05.2006 → 20.05.2006 Konferenznummer: 67762 |
Fingerprint
Untersuchen Sie die Forschungsthemen von „Learning Juntas in the Presence of Noise“. Zusammen bilden sie einen einzigartigen Fingerprint.Projekte
- 1 Abgeschlossen
-
Robuste Lernverfahren und Datenkomprimierung
Reischuk, R. (Projektleiter*in (PI))
01.01.04 → 31.12.08
Projekt: DFG-Projekte › DFG Einzelförderungen