SLIDE 1
Introduction to Artificial Intelligence Deep Learning - Tensor Flow
Janyl Jumadinova December 2, 2016
Credit: Google Workshop
Neural Networks
2/24
Introduction to Artificial Intelligence Deep Learning - Tensor Flow - - PowerPoint PPT Presentation
Introduction to Artificial Intelligence Deep Learning - Tensor Flow - - PowerPoint PPT Presentation
Introduction to Artificial Intelligence Deep Learning - Tensor Flow Janyl Jumadinova December 2, 2016 Credit: Google Workshop Neural Networks 2/24 Neural Networks 3/24 Neural Networks A fully connected NN layer 4/24 Implementation as
SLIDE 2
SLIDE 3
Neural Networks
3/24
SLIDE 4
Neural Networks
A fully connected NN layer
4/24
SLIDE 5
Implementation as Matrix Multiplication
5/24
SLIDE 6
Non-Linear Data Distributions
6/24
SLIDE 7
7/24
SLIDE 8
Deep Learning
◮ Each neuron implements a relatively simple mathematical
function.
◮ y = g(w · x + b) 8/24
SLIDE 9
Deep Learning
◮ Each neuron implements a relatively simple mathematical
function.
◮ y = g(w · x + b) ◮ The composition of 106 − 109 such functions is powerful. 8/24
SLIDE 10
Deep Learning
Book: http://www.deeplearningbook.org/
Chapter 5
“A core idea in deep learning is that we assume that the data was generated by the composition of factors or features, potentially at multiple levels in a hierarchy.”
9/24
SLIDE 11
Results get better with:
◮ more data ◮ bigger models ◮ more computation 10/24
SLIDE 12
Results get better with:
◮ more data ◮ bigger models ◮ more computation
Better algorithms, new insights and improved methods help, too!
10/24
SLIDE 13
11/24
SLIDE 14
Adoption of Deep Learning Tools on GitHub
12/24
SLIDE 15
Tensor Flow
◮ Operates over tensors: n-dimensional arrays 13/24
SLIDE 16
Tensor Flow
◮ Operates over tensors: n-dimensional arrays ◮ Using a flow graph: data flow computation framework 13/24
SLIDE 17
Tensor Flow
◮ Operates over tensors: n-dimensional arrays ◮ Using a flow graph: data flow computation framework 13/24
SLIDE 18
Tensor Flow
◮ 5.7 ← Scalar ◮ Number, Float, etc. 14/24
SLIDE 19
Tensor Flow
15/24
SLIDE 20
Tensor Flow
16/24
SLIDE 21
Tensor Flow
◮ Tensors have a Shape that is described with a vector 17/24
SLIDE 22
Tensor Flow
◮ Tensors have a Shape that is described with a vector ◮ [1000, 256, 256, 3] ◮ 10000 Images ◮ Each Image has 256 Rows ◮ Each Row has 256 Pixels ◮ Each Pixel has 3 values (RGB) 17/24
SLIDE 23
Tensor Flow
Computation is a dataflow graph
18/24
SLIDE 24
Tensor Flow
Computation is a dataflow graph with tensors
19/24
SLIDE 25
Tensor Flow
Computation is a dataflow graph with state
20/24
SLIDE 26
Core TensorFlow data structures and concepts
◮ Graph: A TensorFlow computation, represented as a dataflow
graph:
- collection of ops that may be executed together as a group.
21/24
SLIDE 27
Core TensorFlow data structures and concepts
◮ Graph: A TensorFlow computation, represented as a dataflow
graph:
- collection of ops that may be executed together as a group.
◮ Operation: a graph node that performs computation on tensors 21/24
SLIDE 28
Core TensorFlow data structures and concepts
◮ Graph: A TensorFlow computation, represented as a dataflow
graph:
- collection of ops that may be executed together as a group.
◮ Operation: a graph node that performs computation on tensors ◮ Tensor: a handle to one of the outputs of an Operation:
- provides a means of computing the value in a TensorFlow
Session.
21/24
SLIDE 29
Tensor Flow
◮ Constants 22/24
SLIDE 30
Tensor Flow
◮ Constants ◮ Placeholders: must be fed with data on execution. 22/24
SLIDE 31
Tensor Flow
◮ Constants ◮ Placeholders: must be fed with data on execution. ◮ Variables: a modifiable tensor that lives in TensorFlow’s graph
- f interacting operations.
22/24
SLIDE 32
Tensor Flow
◮ Constants ◮ Placeholders: must be fed with data on execution. ◮ Variables: a modifiable tensor that lives in TensorFlow’s graph
- f interacting operations.
◮ Session: encapsulates the environment in which Operation
- bjects are executed, and Tensor objects are evaluated.
22/24
SLIDE 33
Tensor Flow
23/24
SLIDE 34
Tensor Flow
playground.tensorflow.org
24/24