search

memect / hao

好东西传送门
1.4k stars 459 forks source link

@ lzqkean 有哪些大牛做过GPU加速贝叶斯网络学习，还有相关的代码，方便学习@好东西传送门 #117

Closed haoawesome closed 9 years ago

haoawesome commented 10 years ago

http://www.weibo.com/1974787502/BkfWmyZUV

haoawesome commented 10 years ago

people

https://sites.google.com/site/silbersteinmark/Home Mark Silberstein

Michael Stumpf, Professor for Theoretical Systems Biology at Imperial College London

slides

http://www.slideshare.net/theosysbio/approximate-bayesian-computation-on-gpus Approximate Bayesian Computation on GPUs

http://on-demand.gputechconf.com/gtc/2013/presentations/S3277-Computation-in-Probabilistic-Graphical-Models.pdf

papers

http://arxiv.org/abs/1210.5128 A Novel Learning Algorithm for Bayesian Network and Its Efficient Implementation on GPU (2013)

http://www.jstatsoft.org/v44/i04/paper cudaBayesreg : Parallel Implementation of a Bayesian Multilevel Model for fMRI Data Analysis (2011) Journal of Statistical Software, October 2011, Volume 44, Issue 4.

https://www.ics.uci.edu/~dechter/courses/ics-295/spring-2008/SumProductPaper.pdf Efficient computation of sum-products on GPUs through software-managed cache

code

https://github.com/slinderman/pyglm Generalized linear models for neural spike train modeling, in Python! With GPU-accelerated fully-Bayesian inference, MAP inference, and network priors.

http://gpumlib.sourceforge.net/ GPUMLib is an open source Graphic Processing Unit Machine Learning Library. This library aims to provide machine learning researchers and practitioners with a high performance library by taking advantage of the GPU enormous computational power. The library is developed in C++ and CUDA.

https://github.com/beamandrew/BNN Bayesian Neural Networks

haoawesome commented 10 years ago

http://arxiv.org/abs/1210.5128 A Novel Learning Algorithm for Bayesian Network and Its Efficient Implementation on GPU

haoawesome commented 10 years ago

http://www.reddit.com/r/programming/comments/7j1gr/accelerating_bayesian_network_200x_using_a_gpu/ (2009)

accelerating bayesian network 200X using a GPU , including source code OK, this is making my day a little more surreal, reddit is one of my favorite sites to visit every morning but I never expected to see our group's work show up on the first page. Thanks reddit! For what it's worth, this work was presented this past summer at the International Conference of Supercomputing (on the Greek island of Kos). Mark (first author) led the work while visiting our group at UC Davis in summer 2007.

I'm one of the authors of the paper. I've asked the other authors to visit as well, and as my expertise is the GPU side, I will try to comment on all the GPU questions. Feel free to ask more questions! I'll do my best to answer them.

I also have lots of course material on graphics architecture on last winter's course site: http://www.ece.ucdavis.edu/courses/W08/EEC277/ .

Silberstein et al.(2008) first demonstrated the potential for GPGPU to impact the statistical fitting of simple Bayesian networks,and recent work, such as studies using novel GPU/CPU-based algorithms for MCMC fitting of highly structured Bayesian models of molecular sequence evolution ( Suchard and Rambaut ,2009b, a), clearly exemplifies the opportunities; the latter example realized a greater than 100-fold reduction in run-time in very challenging and otherwise inaccessible computations.

haoawesome commented 10 years ago

是这个问题吗？ accelerating bayesian network 200X using a GPU 问答进展看这里： http://t.cn/RPku09w

http://www.weibo.com/5220650532/BkgNJu9af?ref=home