The goal of this project is to evaluate active learning approaches to learn functions that are costly to evaluate, e.g. physics behaviour, which requires costly simulation or even real-world experiments. Using an active learning approach, the number of function evaluations to acquire training samples should be minimized. Potential candidates for learning approaches are evolutionary algorithms, Baysian methods (e.g. Gaussian mixture regression or Gaussian processes).
Rhythmus | Tag | Uhrzeit | Format / Ort | Zeitraum |
---|
Studiengang/-angebot | Gültigkeit | Variante | Untergliederung | Status | Sem. | LP | |
---|---|---|---|---|---|---|---|
Intelligente Systeme / Master | (Einschreibung bis SoSe 2012) | Projekt | Wahlpflicht | 10 | unbenotet |