A bag-of-features framework for incremental learning of speech invariants in unsegmented audio streams

Olivier Mangin (1) , Pierre-Yves Oudeyer (1) , David Filliat (1, 2)

1 Flowers - Flowing Epigenetic Robots and Systems
2 UEI - Unité d'Électronique et d'informatique

We introduce a computational framework that allows a machine to bootstrap flexible autonomous learning of speech recognition skills. Technically, this framework shall en- able a robot to incrementally learn to recog- nize speech invariants from unsegmented au- dio streams and with no prior knowledge of phonetics. To achieve this, we import the bag-of-words/bag-of-features approach from recent research in computer vision, and adapt it to incremental developmental speech pro- cessing. We evaluate an implementation of this framework on a complex speech database.

Domaines: Robotique [cs.RO]

Olivier Mangin, Pierre-Yves Oudeyer, David Filliat. A bag-of-features framework for incremental learning of speech invariants in unsegmented audio streams. Tenth International Conference on Epigenetic Robotics, 2010, Örenäs Slott, Sweden. ⟨inria-00541802⟩