These postdoc positions focus on how an Embodied Conversational Agent (virtual human) can interact using social language and social nonverbal behaviors, as well as task language and nonverbal behaviors. The main topics of the postdocs are analyzing human-human interaction using hand annotation, online crowdsourcing, and machine learning algorithms, in order to understand how humans use language and nonverbal behavior in social ways, and (b) based on the results of the annotation implementing dialogue modules, using machine learning approaches, that understand the user’s behavior, and generate appropriate social responses in an Embodied Conversational Agent.
The ideal candidates will have complementary backgrounds to one another, in some combination of AI, natural language generation, multi-user architectures, dialogue systems, and machine learning. Experience implementing or working with architectures that handle planning and generation of nonverbal as well as verbal behaviors is preferred. Experience using state-of-the-art approaches to dialogue management, such as deep learning, is a definite plus. The ideal candidates should also be good team players who work well in an exciting and fast-paced multidisciplinary environment.
Skills : Artificial Intelligence, Dialogue Systems, Embodied Conversational Agents, Deep/Machine Learning Position : post-doctoral (potentially 2 post-docs are available with the same description) Duration : 1 year Host Organization : ISIR, Pierre and Marie Curie University, Paris where Justine Cassell will be visiting starting October 15 2017, and continuing through summer 2018. Documents to send : CV, cover letter, names and contact information of 3 people willing to write a letter of recommendation Send documents to : Justine Cassell <justine at cs.cmu.edu>, Catherine Pelachaud<catherine.pelachaud at upmc.fr>, and Mohamed Chetouani <mohamed.chetouani at upmc.fr>
-------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: text/html Size: 4817 bytes Desc: not available URL: <https://mailman.uib.no/public/corpora/attachments/20170727/bdc9ee83/attachment.txt>