Google Cloud Platform Empowers TensorFlow and Machine Learning
Welcome message from author
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Google Cloud Platform Empowers
TensorFlow and Machine Learning
+Kazunori Sato@kazunori_279
Kaz Sato
Staff Developer AdvocateTech Lead for Data & AnalyticsCloud Platform, Google Inc.
What we’ll cover
What is Neural Network and Deep Learning
Machine Intelligence at Google Scale
Cloud Vision API and Speech API
TensorFlow and Cloud Machine Learning
What is Neural Network and Deep Learning
Neural Network is a function that can learn
xn
> b?
w1
wn
x2
x1
Mimicking the behavior of biological neurons
How do you classify them?
weights
bias(threshold)
Programmers would specify the parameters
Let’s see how neural network solves the problem
The computer tries to find the best parameters
A neuron classifies a data point into two kinds
Gradient Descent: changing the params gradually to reduce errors
How do you classify them?
More neurons = More features to extract
Mapping inputs to
a feature space,
classifying with
a hyperplane
From: Neural Networks, Manifolds, and Topology, colah's blog
How about this?
More hidden layers = more hierarchies of features
How about this?
We need to build a Deep Neural Network
From: Convolutional Deep Belief Networks for Scalable Unsupervised Learning of Hierarchical Representations, Honglak Lee et al.
From: mNeuron: A Matlab Plugin to Visualize Neurons from Deep Models, Donglai Wei et. al.
Machine Intelligence at Google scale
The two big challenges of Deep Learning:Computing Power and Training Data
Enterprise
Google Cloud is
The Datacenter as a Computer
Jupiter network
10 GbE x 100 K = 1 Pbps
Consolidates servers with
microsec latency
Borg
No VMs, pure containers
10K - 20K nodes per Cell
DC-scale job scheduling
CPUs, mem, disks and IO
Confidential & ProprietaryGoogle Cloud Platform 26
Google Cloud +
Neural Network =
Google Brain
What's the scalability of Google Brain?
"Large Scale Distributed Systems for Training Neural
Networks", NIPS 2015
○ Inception / ImageNet: 40x with 50 GPUs
○ RankBrain: 300x with 500 nodes
Externalizing the powerto developers
Image analysis with pre-trained models
REST API: receives an image and returns a JSON
No Machine Learning skill required
From $2.50 / 1,000 units (no charge* to try)
General Availability
Cloud Vision API
* You will be charged for Google Cloud Storage and other Google Cloud Platform resources used in your project.
3333
Demo
Pre-trained models. No ML skill required
REST API: receives audio and returns texts
Supports 80+ languages
Streaming or non-streaming
Limited Preview - cloud.google.com/speech
Cloud Speech API
3535
Demo
TensorFlow and Cloud ML
Ready to use Machine Learning models
Use your own data to train models
Cloud Vision API
Cloud Speech API
Cloud Translate API
Cloud Machine Learning
Develop - Model - Test
Google BigQuery
Stay Tuned….
Cloud Storage
Cloud Datalab
NEW
Alpha
GA BetaGA
AlphaGA
GA
Google's open source library for
machine intelligence
tensorflow.org launched in Nov 2015
Used by many production ML projects
What is TensorFlow?
# define the networkimport tensorflow as tfx = tf.placeholder(tf.float32, [None, 784])W = tf.Variable(tf.zeros([784, 10]))b = tf.Variable(tf.zeros([10]))y = tf.nn.softmax(tf.matmul(x, W) + b)
# define a training stepy_ = tf.placeholder(tf.float32, [None, 10])xent = -tf.reduce_sum(y_*tf.log(y))step = tf.train.GradientDescentOptimizer(0.01).minimize(xent)
TensorBoard: visualization tool
Portable and ScalableTraining on:
Mac/Windows
GPU server
GPU cluster / Cloud
Running on:
Android, iOS
RasPi
Distributed Training and Prediction with TensorFlow
Distributed Training with TensorFlow
Data Parallelism
split data,
share model
● CPU/GPU scheduling
● Communications
○ Local, RPC, RDMA
○ 32/16/8 bit quantization
● Cost-based optimization
● Fault tolerance
Distributed Training with TensorFlow
Tensor Processing Unit
ASIC for TensorFlow
Designed by Google
10x better perf / watt
latency and efficiency
bit quantization
TPU on Production
RankBrain
AlphaGo
Google Photos
Speech
and more
Fully managed distributed training and prediction
Supports custom TensorFlow graphs
Integrated with Cloud Dataflow and Cloud Datalab
Limited Preview - cloud.google.com/ml
Cloud Machine Learning (Cloud ML)
Jeff Dean's keynote: YouTube video
Define a custom TensorFlow graph
Training at local: 8.3 hours w/ 1 node
Training at cloud: 32 min w/ 20 nodes (15x faster)
Prediction at cloud at 300 reqs / sec
Cloud ML demo
TensorFlow in the Wild
TensorFlow
powered
Fried Chicken
Nugget Server
From: http://www.rt-net.jp/karaage1/
TensorFlow poweredCucumber Sorter
From: http://workpiles.com/2016/02/tensorflow-cnn-cucumber/
TV popstar classifierwith 95% accuracy
From: http://memo.sugyan.com/entry/2016/06/14/220624
Discriminative Localization
From: https://github.com/jazzsaxmafia/Weakly_detector
Thank you!
Links & Resources
Large Scale Distributed Systems for Training Neural Networks, Jeff Dean and Oriol Vinals
Cloud Vision API: cloud.google.com/vision
Cloud Speech API: cloud.google.com/speech
TensorFlow: tensorflow.org
Cloud Machine Learning: cloud.google.com/ml
Cloud Machine Learning: demo video
Related Documents
