MARCS Afternoon Colloquium (MAC) (Room 3.G.55 - Building 3)
Professor Dan Ventura
Fulbright Senior Scholar; Visiting Professorial Fellow, UNSW Faculty of Art & Design; Professor, Brigham Young University Department of Computer Science
Topic: Hierarchical Bayesian Program Learning as an Approach to Computational Creativity
Speaker: Professor Dan Ventura
Abstract: In Hierarchical Bayesian program learning (HBPL) complex concepts are decomposed into subconcepts, which may in turn be further decomposed. This hierarchical approach assumes some degree of concept factorizablility and leverages that assumption to construct tractable models for subconcepts that can be combined to achieve one-shot or few-shot classification, understanding, and generation for the high-level concept. We suggest that the HBPL framework may be extended beyond simple generation to computational *creativity* in because it allows for explicit modeling of intention, structure, and substructure. We demonstrate this in the domains of microfiction and pop music composition and present working systems—HIEROS and Popstar*—that are built using this approach. We discuss issues related to factoring joint distributions over artefact classes and some ways in which these factors may be modeled. We will also briefly discuss how the factorization of joint distributions may largely reflect the philosophical debates that occur among human creators, suggesting that the HBPL framework might serve as a more precise scaffolding for such debates.
Thank you for your continued attendance at MARCS Monday Afternoon Colloquiums. Let's keep the momentum going.
MARCS staff and students are reminded that all meetings and workshops have an important role in building and maintaining the sense of community which is central to the success of MARCS as a cooperative and energetic research institute. Your attendance is both welcomed and expected.
The zoom ID is: 627 146 998. Link to zoom https://uws.zoom.us/j/627146998