Gergely Neu, an ML researcher at ICREA and Universitat Pompeu Fabra, introduces Value-Driven Transport, a generative modeling framework that integrates optimal control, reinforcement learning, optimal transport, and stochastic primal-dual optimization
Framework evolves samples across five stages via learned value functions.
many of the most successful methods for generative modeling are based on the idea of "measure transport": computing transformations of random variables (sampled from a "source" distribution) to produce samples from a desired "target" distribution. 2/
Gergely Neu@neu_ripsvalue-driven transport! a new framework for generative modeling, combining elements of * optimal control / RL * optimal transport * stochastic primal-dual optimization thread about our new work with @pablorenoz (@UPFBarcelona) & @adrianlmueller (@ETH_en) 1/
@patrickshafto thank you Pat!
Patrick Shafto@patrickshaftoCool
#1426Patrick Shafto@PATRICKSHAFTOCool
Gergely Neu@neu_ripsvalue-driven transport! a new framework for generative modeling, combining elements of * optimal control / RL * optimal transport * stochastic primal-dual optimization thread about our new work with @pablorenoz (@UPFBarcelona) & @adrianlmueller (@ETH_en) 1/