Created: June 09, 2020
Modified: June 09, 2020
Modified: June 09, 2020
variational optimization
This page is from my personal notes, and has not been specifically reviewed for public consumption. It might be incomplete, wrong, outdated, or stupid. Caveat lector.- Holy shit. In December on Galiano I was brainstorming about continuous structure learning and thought of the general trick, for probabilistic programs, and IIRC thought about a general trick of converting an optimization over discrete program traces (?) into optimizing the mixture weights of a variational distribution over program traces
- I think that this is the same idea as in this paper: Staines and Barber, 2012. Variational Optimization. https://arxiv.org/abs/1212.4507 (papers to read)
- TODO: try to recover my ideas. read the paper. understand if:
- I reinvented the paper (cool!).
- or my case is importantly different from the paper (good to understand).
- If I did reinvent the paper, then is the probabilistic-program perspective on it at least novel and interesting?