[Adapt] [Seminar] Probabilistic Programming on Spark

[Zhao Zhuoyue]

Hi ADAPTers,

Our next siminar will be tomorrow at 5pm in room 3-528. I'll introduce our ongoing project "Probabilistic Programming on Spark". We submitted a visonary paper to SIGMOD Record recently.

Probabilistic programming languages simplify the definition of probabilistic models and perform automatic inference whenever possible. Existing probabilistic programming languages cannot easily scale up to very large distributed data set. Our proposal is to extend Scala, a object-oriented functional programming language on which Spark is built, with probabilistic constructs. The inference code can utilize Spark, an in-memory large-scale mapreduce library, to enable large-scale machine learning tasks. In this talk, I'll introduce a classic general inference algorithm, Variational Message Passing, one of the main inference algorithms we'll use in our project. Hope you'll enjoy the talk.

Thanks,


Zhao Zhuoyue
ACM Class, Shanghai Jiao Tong University
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://cs.sjtu.edu.cn/pipermail/adapt/attachments/20150428/8a0b75eb/attachment-0001.html>