Successfully reported this slideshow.
We use your LinkedIn profile and activity data to personalize ads and to show you more relevant ads. You can change your ad preferences anytime.

確率的プログラミングライブラリEdward

12,163 views

Published on

@Google Cloud Next'17 Cloud Community fes

Published in: Software
  • Hello! Get Your Professional Job-Winning Resume Here - Check our website! https://vk.cc/818RFv
       Reply 
    Are you sure you want to  Yes  No
    Your message goes here
  • Download this 3-step guide to generating insane amounts of media coverage for your from LinkedIn: http://bit.ly/linkedin3stepguide
       Reply 
    Are you sure you want to  Yes  No
    Your message goes here

確率的プログラミングライブラリEdward

  1. 1. Edward 2017-06-14 Cloud Community fes @ Google Cloud Next'17 Tokyo
  2. 2. Yuta Kashino ( ) BakFoo, Inc. CEO Astro Physics /Observational Cosmology Zope / Python Realtime Data Platform for Enterprise / Prototyping
  3. 3. Yuta Kashino ( ) arXiv stat.ML, stat.TH, cs.CV, cs.CL, cs.LG math-ph, astro-ph PyCon2016 @yutakashino https://www.slideshare.net/yutakashino/pyconjp2016
  4. 4. Edward
  5. 5. Edward - Dustin Tran (Open AI) - Blei Lab - (PPL) - Stan, PyMC3, Anglican, Church, Venture,Figaro, WebPPL - 2016 2 PPL - TensorFlow - George Edward Pelham Box Box-Cox Trans., Box-Jenkins, Ljung-Box test box plot Tukey, 3 2 RA Fisher
  6. 6. PPL Edward TensorFlow(TF) + (PPL) TF: PPL: + + Python/Numpy
  7. 7. 1. TF:
  8. 8. 1. TF: - - :
  9. 9. 1. TF:
  10. 10. 1. TF: - - - GPU / TPU Inception v3 Inception v4
  11. 11. 1. TF: - Keras, Slim - TensorBoard
  12. 12. 1. TF: - DNN: NN - - … - = : DropOut DropOut : Yingzhen Li, Yarin Gal ICML, 2017
  13. 13. 1. TF: - = - TF - Edward
  14. 14. 2.
  15. 15. 2. x: edward x⇤ s P(x | ↵) ✓⇤ ⇠ Beta(✓ | 1, 1)
  16. 16. 2. - ( ) Edward p(x, ✓) = Beta(✓ | 1, 1) 50Y n=1 Bernoulli(xn | ✓),
  17. 17. 2. - log_prob() - mean() - sample()
  18. 18. 3.
  19. 19. 3. Edward TF
  20. 20. 3. 256 28*28
  21. 21. 4.
  22. 22. 4. X, Z Z - (Variational Bayes) - (MCMC) p(z | x) = p(x, z) R p(x, z)dz .
  23. 23. 4.
  24. 24. 4. p(z|x) KL q(z) ELBO
  25. 25. 4. Edward KLqp
  26. 26. 5. Box’s loop
  27. 27. 5. Box’s loop George Edward Pelham Box Blei 2014
  28. 28. 5. Box’s loop
  29. 29. Edward - Edward = TensorFlow + + - TensorFlow - - TF GPU, TPU, TensorBoard, Keras - - Box’s Loop - Python
  30. 30. Refrence •D. Tran, A. Kucukelbir, A. Dieng, M. Rudolph, D. Liang, and D.M. Blei. Edward: A library for probabilistic modeling, inference, and criticism.(arXiv preprint arXiv:1610.09787) •D. Tran, M.D. Hoffman, R.A. Saurous, E. Brevdo, K. Murphy, and D.M. Blei. Deep probabilistic programming.(arXiv preprint arXiv:1701.03757) •Box, G. E. (1976). Science and statistics. (Journal of the American Statistical Association, 71(356), 791–799.) •D.M. Blei. Build, Compute, Critique, Repeat: Data Analysis with Latent Variable Models. (Annual Review of Statistics and Its Application Volume 1, 2014)
  31. 31. Questions kashino@bakfoo.com @yutakashino
  32. 32. BakFoo, Inc. NHK NMAPS: +
  33. 33. BakFoo, Inc. PyConJP 2015 Python
  34. 34. BakFoo, Inc.
  35. 35. BakFoo, Inc. : SNS +

×