• Martin Thoma
  • Home
  • Categories
  • Tags
  • Archives
  • Support me

ML Review 5

Contents

  • ML Review 5
    • News
    • Live Demos, Websites and Blogs
    • Publications
    • Software
    • Interesting Questions
    • Miscallenious
      • Blogs / Websites
    • Meetings

This Review gives an overview of intersting stuff I stumbled over which are related to machine learning.

News

  • Canada’s AI Moment
  • AI is one step closer to mastering StarCraft: Multiagent Bidirectionally-Coordinated Nets for Learning to Play StarCraft Combat Game
  • Deep learning for satellite imagery via image segmentation
  • Hardware:
    • Nvidia Volta (reddit)
    • Nvidia BB 8 (self-driving car)
    • First In-Depth Look at Google’s TPU Architecture
  • Datasets:
    • Medical ImageNet
    • READ-BAD - detecting baselines in scanned documents
  • Google’s AlphaGo AI still undefeated, beats world’s best human Ke Jie again

Live Demos, Websites and Blogs

  • Jonathan Pilault: Adam-Nesterov results: Related to the OpenAI gym.
  • lyrebird.ai: Imitate voices with only one minute of recorded speech

Publications

  • Beyond Face Rotation: Global and Local Perception GAN for Photorealistic and Identity Preserving Frontal View Synthesis - generate frontal view faces from side view input
  • Cyclical Learning Rates for Training Neural Networks
  • BAM! The Behance Artistic Media Dataset for Recognition Beyond Photography: A dataset with 393000 labeled images, 7 types of images (vector art, pencil-sketch, ...) and 9 classes

Software

Text
FaceApp
  • caffe2 - although caffe will be continued, it is probably a good idea to switch?
  • CycleGAN: Translate between horses and Zebras
  • FaceApp: See this and reddit
  • fastText: library for efficient learning of word representations and sentence classification
  • FBLearner Flow

Interesting Questions

  • What are 2D dimensionality reduction algorithms good for?
  • What is the memory cost of a CNN?
  • How should the bias be initialized and regularized?
  • Is there a relationship between LDA, linear SVMs and Perceptron?

Miscallenious

Blogs / Websites

  • distill.pub: Some nice articles (e.g. one about t-SNE)

Meetings

  • Amsterdam, 8. April 2017: PyData
  • Austin (Texas, USA), 10. July 2017: SciPy 2017
« Previous Review

Published

Mai 25, 2017
by Martin Thoma

Category

Machine Learning

Tags

  • Machine Learning 81
  • matplotlib 7

Contact

  • Martin Thoma - A blog about Code, the Web and Cyberculture
  • E-mail subscription
  • RSS-Feed
  • Privacy/Datenschutzerklärung
  • Impressum
  • Powered by Pelican. Theme: Elegant by Talha Mansoor