# DL Classe 1 - Go Deep or Go Home

Post on 16-Apr-2017

78 views

Embed Size (px)

TRANSCRIPT

Louis Monier@louis_monier

https://www.linkedin.com/in/louismonier

Deep Learning BasicsGo Deep or Go Home

Gregory Renard@redohttps://www.linkedin.com/in/gregoryrenard

Class 1 - Q1 - 2016

Supervised Learning

cathatdogcatmug

=?

prediction

dog (89%)

training set

Training(days)

Testing(ms)

Single Neuron to Neural Networks

Single Neuron

+

x

w1

xpos

x

w2

ypos

w3

ChoiceScore

Linear Combination Nonlinearity

3.4906 0.9981, red

Inputs

3.45

1.21

-0.314

1.59

2.65

More Neurons, More Layers: Deep Learning

input1

input2

input3

input layer hidden layers output layer

circleiris

eye

face

89% likely to be a face

How to find the best values for the weights?

w = Parameter

Error

a

b

Define error = |expected - computed| 2

Find parameters that minimize average error

Gradient of error wrt w is

Update:

take a step downhill

Egg carton in 1 million dimensions

Get to here

or here

Backpropagation: Assigning Blame

input1

input2

input3

label=1

loss=0.49

w1

w2

w3

b

w1

w2

w3

b

1). For loss to go down by 0.1, p must go up by 0.05.For p to go up by 0.05, w1 must go up by 0.09.For p to go up by 0.05, w2 must go down by 0.12....

prediction p=0.3

p=0.89

2). For loss to go down by 0.1, p must go down by 0.01

3). For p to go down by 0.01, w2 must go up by 0.04

Stochastic Gradient Descent (SGD) and Minibatch

Its too expensive to compute the gradient on all inputs to take a step.

We prefer a quick approximation.

Use a small random sample of inputs (minibatch) to compute the gradient.

Apply the gradient to all the weights.

Welcome to SGD, much more efficient than regular GD!

Usually things would get intense at this point...

Chain rule

Partial derivatives

Hessian

Sum over paths

Credit: Richard Socher Stanford class cs224d

Jacobian matrix

Local gradient

Learning a new representation

Credit: http://colah.github.io/posts/2014-03-NN-Manifolds-Topology/

Training Data

Train. Validate. Test.Training set

Validation set

Cross-validation

Testing set

2 3 41

2 3 41

2 3 41

2 3 41

shuffle!

one epoch

Never ever, ever, ever, ever, ever, ever use for training!

Tensors

Tensors are not scary

Number: 0D

Vector: 1D

Matrix: 2D

Tensor: 3D, 4D, array of numbers

This image is a 3264 x 2448 x 3 tensor

[r=0.45 g=0.84 b=0.76]

Different types of networks / layers

Fully-Connected NetworksPer layer, every input connects to every neuron.

input1

input2

input3

input layer hidden layers output layer

circleiris

eye

face

89% likely to be a face

Recurrent Neural Networks (RNN)Appropriate when inputs are sequences.

Well cover in detail when we work on text.

pipe a not is this

t=0t=1t=2t=3t=4

this is not a pipe

t=0 t=1 t=2 t=3 t=4

Convolutional Neural Networks (ConvNets)Appropriate for image tasks, but not limited to image tasks.

Well cover in detail in the next session.

HyperparametersActivations: a zoo of nonlinear functions

Initializations: Distribution of initial weights. Not all zeros.

Optimizers: driving the gradient descent

Objectives: comparing a prediction to the truth

Regularizers: forcing the function we learn to remain simple

...and many more

ActivationsNonlinear functions

Sigmoid, tanh, hard tanh and softsign

ReLU (Rectified Linear Unit) and Leaky ReLU

Softplus and Exponential Linear Unit (ELU)

Softmax

x1 = 6.78

x2 = 4.25

x3 = -0.16

x4 = -3.5

dog = 0.92536

cat = 0.07371

car = 0.00089

hat = 0.00003

Raw scores Probabilities

SOFTMAX

Optimizers

Various algorithms for driving Gradient DescentTricks to speed up SGD.

Learn the learning rate.

Credit: Alex Radford

Cost/Loss/Objective Functions

=?

Cross-entropy Loss

x1 = 6.78

x2 = 4.25

x3 = -0.16

x4 = -3.5

dog = 0.92536

cat = 0.07371

car = 0.00089

hat = 0.00003

Raw scores Probabilities

SOFTMAX

dog: loss = 0.078

cat: loss = 2.607

car: loss = 7.024

hat: loss = 10.414

If label is:

=?

label (truth)

RegularizationPreventing Overfitting

Overfitting

Very simple.Might not predict well.

Just right? Overfitting. Rote memorization.Will not generalize well.

Regularization: Avoiding Overfitting

Idea: keep the functions simple by constraining the weights.

Loss = Error(prediction, truth) + L(keep solution simple)

L2 makes all parameters medium-size

L1 kills many parameters.

L1+L2 sometimes used.

At every step, during training, ignore the output of a fraction p of neurons.

p=0.5 is a good default.

Regularization: Dropout

input1

input2

input3

One Last Bit of Wisdom

Workshop : Keras & Addition RNNhttps://github.com/holbertonschool/deep-learning/tree/master/Class%20%230

Recommended