On Friday October 17, this site was moved to a new server, https://mw.hh.se. The original address will continue to work. Whithin a week or two this site will return to the original address. /Peo HH IT-dep
WG211/M17Shan: Difference between revisions
From WG 2.11
Jump to navigationJump to search
Chung-chieh (talk | contribs) Abstract |
Chung-chieh (talk | contribs) mNo edit summary |
||
| Line 1: | Line 1: | ||
'''Inference building blocks''' by Chung-chieh Shan | '''Inference building blocks''' by Chung-chieh Shan | ||
(I don't have to speak; I' | (I don't have to speak; I'll only be around on July 17 and 18) | ||
How can we make probabilistic inference techniques easier to compose, out of building blocks that comprise an inference language? Our answer to this question is to develop over the last few years an arsenal of automatic transformations whose inputs and outputs are programs in a unifying language of distributions. By expressing inference techniques in terms of these transformations, we can mix and match them with a family of similar probabilistic models and achieve practical performance. | How can we make probabilistic inference techniques easier to compose, out of building blocks that comprise an inference language? Our answer to this question is to develop over the last few years an arsenal of automatic transformations whose inputs and outputs are programs in a unifying language of distributions. By expressing inference techniques in terms of these transformations, we can mix and match them with a family of similar probabilistic models and achieve practical performance. | ||
Latest revision as of 01:57, 30 June 2017
Inference building blocks by Chung-chieh Shan
(I don't have to speak; I'll only be around on July 17 and 18)
How can we make probabilistic inference techniques easier to compose, out of building blocks that comprise an inference language? Our answer to this question is to develop over the last few years an arsenal of automatic transformations whose inputs and outputs are programs in a unifying language of distributions. By expressing inference techniques in terms of these transformations, we can mix and match them with a family of similar probabilistic models and achieve practical performance.
