README.md

MindSpore MDP Special Interest Group (SIG)

This is the working repo for the MDP Special Interest Group (SIG). MindSpore Deep Probabilistic Programming (MDP) is a programming library for Bayesian deep learning. The target of MDP is to intergrade the gap between deep learning and Bayesian learning. This repo contains all the artifacts, materials, meeting notes and proposals regarding Probabilistic Programming , Deep Probabilistic Programming , Toolbox . Feedbacks and contributions are welcomed.

1. Probabilistic Programming: Probabilistic Programming (PP) focuses on professional Bayesian learning, including statistical distributions classes used to generate stochastic tensors and probabilistic inference algorithms.
2. Deep Probabilistic Programming: Deep Probabilistic Programming (DPP) aims to provide composable BNN modules, which contains bnn layers, bnn modules, transforms and context.
3. Toolbox: Toolbox provides a set of BNN tools for some specific applications, such as Uncertainty Estimation, OoD Detection and so on.

SIG Leads

Chen Jianfei (Tsinghua University)

Logistics

• SIG leads will drive the meeting.
• Meeting announcement will be posted on our gitee channel: https://gitee.com/mindspore/community/tree/master/sigs/mdp
• Feedbacks and topic requests are welcomed by all.

Discussion

• Slack channel: https://app.slack.com/client/T018BLCMSGL/learning-slack
• Documents and artifacts: https://gitee.com/mindspore/community/tree/master/sigs/mdp

Meeting notes