Concentration lesson that learns deep learning (deep learning) with slide and movie in three hours made by Google developers

Machine learning using neural network "Deep learning (deep learning)Is an indispensable technology for artificial intelligence (AI) development, and it is utilized in technologies that have already been put into practical use, such as automatic photo tagging function of Google photo image and recommendation function of Amazon in addition to AI. Google's cloud developer introduces the best way for developers to learn such deep learning by blogging.

Learn Tensor Flow and deep learning, without a Ph.D. | Google Cloud Big Data and Machine Learning Blog | Google Cloud Platform
https://cloud.google.com/blog/big-data/2017/01/learn-tensorflow-and-deep-learning-without-a-phd

Martin Gorner, who develops the cloud platform on Google, said he felt that "the information overflowing on the net is too difficult for students to learn about deep learning, which can not be learned." In the information for developers who explain deep learning, as a precondition to understand explanation "Cross entropy"The esoteric concept such as" disappearance of gradient "is used as it is, so it is a content that can not be understood by outsider experts.

Mr. Gorner who remembered the university professor's reprimand that "Integral has learned in kindergarten" in a situation where content which is too much difficulty is full is deeply developed by developers without acquiring the Ph.D. I decided to develop a way to learn the learning technique. Mr. Gorner developed a concentration lesson that learns the mechanism of Google's machine learning library "TensorFlow" and deep learning in total of 3 hours. In this intensive lesson, "Convolution neural network"Recursive neural networkIt seems that it focuses on the basic network structure such as "

"3 hour intensive lesson to understand development method of deep learning" developed by Mr. Gorner was held in Belgium in November 2016Devoxx conferenceIt was announced as a presentation. You can check the state of the intensive lesson with the following movies.

Tensorflow and deep learning - without a PhD by Martin Görner - YouTube


The slides used in this intensive lesson can be viewed on the following sites.

Tensorflow and deep learning - without a PhD - Google slide
https://docs.google.com/presentation/d/1TVixw6ItiZ8igjp6U17tcgoFrLSaHWQmMOwjlgQY9co/


Tensorflow and deep learning 2 - without a PhD - Google Slide
https://docs.google.com/presentation/d/e/2PACX-1vRouwj_3cYsmLrNNI3Uq5gv5-hYp_QFdeoan2GlxKgIZRSejozruAbVV0IMXBoPsINB7Jw


The learning items are as follows. Neural Network and TensorFlow Basics


Multiple layers, ReLU, Dropout, Decay rate


Convolution neural network


Batch normalization


High level API


Recursive neural network


Google cloud machine learning platform


There is also an hour version for the busy people "I can not use 3 hours, make it an hour!"

Tensorflow and deep learning - without a PhD by Martin Görner - YouTube


It might be a good idea to try a concentration lesson for 3 hours to 1 hour first before being overwhelmed by the advanced parade on parade trying to learn deep learning.