JOINT POSTDOC POSITION IN MACHINE LEARNING AT THE ECOLE NORMALE SUPERIEURE …

The general aim of this project is to explore unsupervised learning algorithms for speech inspired by how human infants spontaneously learn their first language through mere immersion in a language speaking community.Specifically, the research will focus on discovering speaker and context invariant lexical or sublexical (phoneme-like) units in an unsupervised fashion from raw speech. This will done within the framework of the zero resource speech challenge (www.zerospeech.com, [1]), and will use various machine learning techniques (pattern matching, dictionary learning methods, generative Bayesian models), including methods used in vision, like epitomic representations [2].The post-doc will work in a multidisciplinary team composed of researchers with various backgrounds (neuroscience, psycholinguistics, machine learning, etc) bi-located (Ecole Normale Supérieure in the quartier latin in Paris, Microsoft Research Labs at Redmond, Seattle), and will have access…


Link to Full Article: JOINT POSTDOC POSITION IN MACHINE LEARNING AT THE ECOLE NORMALE SUPERIEURE …