Deep Learning Lab Course 2017

(Deep Learning Practical)

Labs:
(Computer Vision) Thomas Brox,
(Robotics) Wolfram Burgard,
(Machine Learning) Frank Hutter,
(Neurorobotics) Joschka Boedecker

University of Freiburg

January 12, 2018

Andreas Eitel Uni FR DL 2017 (1)

 Technical Issues

I Location: Monday, 14:00 - 16:00, building 082, room 00 028

I Remark: We will be there for questions every week from 14:00 - 16:00.
I We expect you to work on your own. Your attendance is required during
lectures/presentations
I We expect you have basic knowledge in ML (e.g. heard the Machine
Learning lecture).
I Contact information (tutors):
Aaron Klein [email protected]
Artemij Amiranashvili [email protected]
Mohammadreza Zolfaghari [email protected]
Maria Hügle [email protected]
Jingwhei Zhang [email protected]
Andreas Eitel [email protected]
I Homepage:
http://ais.informatik.uni-freiburg.de/teaching/ws17/deep learning course

Andreas Eitel Uni FR DL 2017 (2)

 Schedule and outline

I Phase 1
I Today: Introduction to Deep Learning (lecture).
I Assignment 1) 16.10 - 06.11 Implementation of a feed-forward Neural
Network from scratch (each person individually, 1-2 page report).
I 23.10: meeting solving open questions
I 30.10: meeting solving open questions
I 06.11: Intro to CNNs and Tensorflow (mandatory lecture), hand in
Assignment 1
I Assignment 2) 06.11 - 20.11 Train CNN in tensorflow (each person
individually, 1-2 page report)
I 13.11 meeting solving open questions

Andreas Eitel Uni FR DL 2017 (3)

 Schedule and outline

I Phase 1
I Today: Introduction to Deep Learning (lecture).
I Assignment 1) 16.10 - 06.11 Implementation of a feed-forward Neural
Network from scratch (each person individually, 1-2 page report).
I 23.10: meeting solving open questions
I 30.10: meeting solving open questions
I 06.11: Intro to CNNs and Tensorflow (mandatory lecture), hand in
Assignment 1
I Assignment 2) 06.11 - 20.11 Train CNN in tensorflow (each person
individually, 1-2 page report)
I 13.11 meeting solving open questions
I Phase 2 (split into three tracks)
I 20.11 First meeting in tracks, hand in Assignment 2
I Assignment 3) 20.11 - 18.12 Group work track specific task, 1-2 page
report
I Assignment 4) 18.12 - 22.01 Group work advanced topic, 1-2 page report

Andreas Eitel Uni FR DL 2017 (3)

 Schedule and outline

I Phase 1
I Today: Introduction to Deep Learning (lecture).
I Assignment 1) 16.10 - 06.11 Implementation of a feed-forward Neural
Network from scratch (each person individually, 1-2 page report).
I 23.10: meeting solving open questions
I 30.10: meeting solving open questions
I 06.11: Intro to CNNs and Tensorflow (mandatory lecture), hand in
Assignment 1
I Assignment 2) 06.11 - 20.11 Train CNN in tensorflow (each person
individually, 1-2 page report)
I 13.11 meeting solving open questions
I Phase 2 (split into three tracks)
I 20.11 First meeting in tracks, hand in Assignment 2
I Assignment 3) 20.11 - 18.12 Group work track specific task, 1-2 page
report
I Assignment 4) 18.12 - 22.01 Group work advanced topic, 1-2 page report
I Phase 3:
I 15.01: meeting final project topics and solving open questions
I Assignment 5) 22.01 - 12.02 Solve a challenging problem of your choice
using the tools from Phase 2 (group work + 5 min presentation)

Andreas Eitel Uni FR DL 2017 (3)

 Tracks

I Track 1 Neurorobotics/Robotics
I Robot navigation
I Deep reinforcement learning

Andreas Eitel Uni FR DL 2017 (4)

 Tracks

I Track 1 Neurorobotics/Robotics
I Robot navigation
I Deep reinforcement learning
I Track 2 Machine Learning
I Architecture search and hyperparameter optimization
I Visual recognition

Andreas Eitel Uni FR DL 2017 (4)

 Tracks

I Track 1 Neurorobotics/Robotics
I Robot navigation
I Deep reinforcement learning
I Track 2 Machine Learning
I Architecture search and hyperparameter optimization
I Visual recognition
I Track 3 Computer Vision
I Image segmentation
I Autoencoders
I Generative adversarial networks

Andreas Eitel Uni FR DL 2017 (4)

 Groups and Evaluation

I In total we have three practical phases

I Reports:
I Short 1-2 page report explaining your results, typically accompanied by 1-2
figures (e.g. a learning curve)
I hand in the code you used to generate these results as well, send us your
github/bitbucket repo
I Presentations:
I Phase 3: 5 min group work presentation in last week
I the reports and presentation after each exercise will be evaluated w.r.t.
scientific quality and quality of presentation
I Requirements for passing:
I attending the class in the lecture and presentation weeks
I active participation in the small groups, i.e., programming, testing, writing
report

Andreas Eitel Uni FR DL 2017 (5)

 Today...

I Groups: You will work on your own for the first exercise
I Assignment: You will have three weeks to build a MLP and apply it to
the MNIST dataset (more on this at the end)
I Lecture: Short recap on how MLPs (feed-forward neural networks) work
and how to train them

Andreas Eitel Uni FR DL 2017 (6)

 (Deep) Machine Learning Overview

Data
?
inference

Model M
1 Learn Model M from the data
2 Let the model M infer unknown quantities from data

Andreas Eitel Uni FR DL 2017 (7)

 (Deep) Machine Learning Overview

Data
?
inference

Model M

Andreas Eitel Uni FR DL 2017 (8)

 Machine Learning Overview

What is the difference between deep learning and a standard machine learning
pipeline ?

Andreas Eitel Uni FR DL 2017 (9)

 Standard Machine Learning Pipeline

(1) Engineer good features (not learned)

(2) Learn Model
(3) Inference e.g. classes of unseen data

x 1

x 2

x
(3)
3

features
(1) x 1 color height (4 cm) (2)
x 2
color height (3.5 cm)

x 3
color height (10 cm)

Data

Andreas Eitel Uni FR DL 2017 (10)

 Unsupervised Feature Learning Pipeline

(1a) Maybe engineer good features (not learned)

(1b) Learn (deep) representation unsupervisedly
(2) Learn Model
(3) Inference e.g. classes of unseen data

x 1

x 2

features
x x 1 color height (4 cm) (3)
3
x 2
color height (3.5 cm)

x
(1a) 3
color height (10 cm)

(2)

Data Pixels

(1b)
Andreas Eitel Uni FR DL 2017 (11)
 Supervised Deep Learning Pipeline

(1) Jointly Learn everything with a deep architecture

(2) Inference e.g. classes of unseen data

x 1
Classes

x 2

(2)
x 3

Pixels

Data
(1)

Andreas Eitel Uni FR DL 2017 (12)

 Training supervised feed-forward neural networks

I Let’s formalize!
I We are given:
I Dataset D = {(x1 , y1 ), . . . , (xN , yN )}
I A neural network with parameters θ which implements a function fθ (x)
I We want to learn:
I The parameters θ such that ∀i ∈ [1, N ] : fθ (xi ) = yi

Andreas Eitel Uni FR DL 2017 (13)

 Training supervised feed-forward neural networks

I A neural network with parameters θ which implements a function fθ (x)

→ θ is given by the network weights w and bias terms b
smax smax
f(x) a a
1
(2)
h (x) a
... a
... a
1
(1)
h (x) a
... a
... ai
1

x ... x ... x
1 j N

Andreas Eitel Uni FR DL 2017 (14)

 Neural network forward-pass

I Computing fθ (x) for a neural network is a forward-pass

smax smax
f(x) a a
1
I unit i activation:
XN
(2)
h (x) a
... a
... a
ai = wi,j xj + bi
j=i
1
unit i output:
... ...
I (1)
(1) h (x) a a ai
hi (x) = t(ai ) where t(·) is an
activation or transfer function b i

w i,j
1

x ... x ... x
1 j N

Andreas Eitel Uni FR DL 2017 (15)

 Neural network forward-pass

I Computing fθ (x) for a neural network is a forward-pass

smax smax
f(x) a a
alternatively (and much faster) use 1
vector notation: (2)
h (x) a
... a
... a
I layer activation:
a(1) = W(1) x + b(1)
1
I layer output:
h(1) (x) = t(a(1) )
(1)
h (x) a
... a
... ai
where t(·) is applied element wise (1) b i

W w i,j
1

x ... x ... x
1 j N

Andreas Eitel Uni FR DL 2017 (16)

 Neural network forward-pass

I Computing fθ (x) for a neural network is a forward-pass

smax smax
f(x) a a
Second layer 1

I layer 2 activation:
(2)
h (x) a
... a
... a
a(2) = W(2) h(1) (x) + b(2) (2)
I layer 2 output: W 1
h(1) (x) = t(a(2) )
(1)
h (x) a
... a
... ai
where t(·) is applied element wise
(1)
W 1

x ... x ... x
1 j N

Andreas Eitel Uni FR DL 2017 (17)

 Neural network forward-pass

I Computing fθ (x) for a neural network is a forward-pass

smax smax
Output layer f(x) a a
output layer activation: (3)
W
I
1
a(3) = W(3) h(2) (x) + b(3)
I network output:
(2)
h (x) a
... a
... a
f (x) = o(a(3) ) (2)
where o(·) is the output W 1
nonlinearity (1)
h (x) a
... a
... ai
I for classification use softmax:
ezi (1)
oi (z) = P|z|
j=1 e
zj W 1

x ... x ... x
1 j N

Andreas Eitel Uni FR DL 2017 (18)

 Training supervised feed-forward neural networks

Neural network activation functions

I

Typical nonlinearities for hidden layers are: tanh(ai ),

I
1
sigmoid σ(ai ) = , ReLU relu(ai ) = max(ai , 0)
1 + e−ai
I tanh and sigmoid are both squashing non-linearities
I ReLU just thresholds at 0

→ Why not linear ?

Andreas Eitel Uni FR DL 2017 (19)

 Training supervised feed-forward neural networks

I Train parameters θ such that ∀i ∈ [1, N ] : fθ (xi ) = yi

Andreas Eitel Uni FR DL 2017 (20)

 Training supervised feed-forward neural networks

I Train parameters θ such that ∀i ∈ [1, N ] : fθ (xi ) = yi

I We can do this via minimizing the empirical risk on our dataset D
N
1 X
min L(fθ , D) = min l(fθ (xi ), yi ), (1)
θ θ N i=1

where l(·, ·) is a per example loss

Andreas Eitel Uni FR DL 2017 (20)

 Training supervised feed-forward neural networks

I Train parameters θ such that ∀i ∈ [1, N ] : fθ (xi ) = yi

I We can do this via minimizing the empirical risk on our dataset D
N
1 X
min L(fθ , D) = min l(fθ (xi ), yi ), (1)
θ θ N i=1

where l(·, ·) is a per example loss

I For regression often use the squared loss:
M
1X
l(fθ (x), y) = (fj,θ (x) − yj )2
2 j=1

I For M-class classification use the negative log likelihood:

M
X
l(fθ (x), y) = −log(fj,θ (x)) · yj
j

Andreas Eitel Uni FR DL 2017 (20)

 Gradient descent

I The simplest approach to minimizing min L(fθ , D) is gradient descent

θ

Gradient descent:
θ0 ← init randomly
do
t+1 ∂L(fθ , D)
I θ = θt − γ
∂θ
while
(L(fθt+1 , V ) − L(fθt , V ))2 >


Andreas Eitel Uni FR DL 2017 (21)

 Gradient descent

I The simplest approach to minimizing min L(fθ , D) is gradient descent

θ

Gradient descent:
θ0 ← init randomly
do
t+1 ∂L(fθ , D)
I θ = θt − γ
∂θ
while
(L(fθt+1 , V ) − L(fθt , V ))2 >

I Where V is a validation dataset (why not use D ?)
N
1 X
I Remember in our case: L(fθ , D) = l(fθ (xi ), yi )
N i=1
I We will get to computing the derivatives shortly

Andreas Eitel Uni FR DL 2017 (21)

 Gradient descent

I Gradient descent example: D = {(x1 , y 1 ), . . . , (x100 , y 100 )} with

x ∼ U [0, 1]
y = 3 · x +  where  ∼ N (0, 0.1)

Andreas Eitel Uni FR DL 2017 (22)

 Gradient descent

I Gradient descent example: D = {(x1 , y 1 ), . . . , (x100 , y 100 )} with

x ∼ U [0, 1]
y = 3 · x +  where  ∼ N (0, 0.1)

I Learn parameters θ of function fθ (x) = θx using loss

1 1
l(fθ (x), y) = kfθ (x) − yk22 = (fθ (x) − y)2
2 2
N N
∂L(fθ , D) 1 X ∂l(fθ , D) 1 X
= = (θx − y)x
∂θ N i=1 ∂θ N i=1

Andreas Eitel Uni FR DL 2017 (22)

 Gradient descent

I Gradient descent example: D = {(x1 , y 1 ), . . . , (x100 , y 100 )} with

x ∼ U [0, 1]
y = 3 · x +  where  ∼ N (0, 0.1)

I Learn parameters θ of function fθ (x) = θx using loss

1 1
l(fθ (x), y) = kfθ (x) − yk22 = (fθ (x) − y)2
2 2
N N
∂L(fθ , D) 1 X ∂l(fθ , D) 1 X
= = (θx − y)x
∂θ N i=1 ∂θ N i=1

Andreas Eitel Uni FR DL 2017 (22)

 Gradient descent

I Gradient descent example γ = 2.

gradient descent

Andreas Eitel Uni FR DL 2017 (23)

 Stochastic Gradient descent (SGD)

I There are two problems with gradient descent:

1. We have to find a good γ
2. Computing the gradient is expensive if the training dataset is large!
I Problem 2 can be attacked with online optimization
(we will have a look at this)
I Problem 1 remains but can be tackled via second order methods or other
advanced optimization algorithms (rprop/rmsprop, adagrad)

Andreas Eitel Uni FR DL 2017 (24)

 Gradient descent

1. We have to find a good γ (γ = 2., γ = 5.)

gradient descent 2

Andreas Eitel Uni FR DL 2017 (25)

 Stochastic Gradient descent (SGD)

2 Computing the gradient is expensive if the training dataset is large!

I What if we would only evaluate f on parts of the data ?

Stochastic Gradient Descent:

θ0 ← init randomly
do
I (x0 , y0 ) ∼ D
sample example from dataset D
t+1 ∂l(fθ (x0 ), y0 )
I θ = θt − γ t
∂θ
while (L(fθt+1 , V ) − L(fθt , V ))2 > 
∞
X X∞
where γ t → ∞ and (γ t )2 < ∞
t=1 t=1
(γ should go to zero but not too fast)

Andreas Eitel Uni FR DL 2017 (26)

 Stochastic Gradient descent (SGD)

2 Computing the gradient is expensive if the training dataset is large!

I What if we would only evaluate f on parts of the data ?

Stochastic Gradient Descent:

θ0 ← init randomly
do
I (x0 , y0 ) ∼ D
sample example from dataset D
t+1 ∂l(fθ (x0 ), y0 )
I θ = θt − γ t
∂θ
while (L(fθt+1 , V ) − L(fθt , V ))2 > 
∞
X X∞
where γ t → ∞ and (γ t )2 < ∞
t=1 t=1
(γ should go to zero but not too fast)
→ SGD can speed up optimization for large datasets
→ but can yield very noisy updates
→ in practice mini-batches are used
(compute l(·, ·) for several samples and average)
→ we still have to find a learning rate shedule γ t
Andreas Eitel Uni FR DL 2017 (26)
 Stochastic Gradient descent (SGD)

→ Same data, assuming that gradient evaluation on all data takes 4 times as
much time as evaluating a single datapoint
1
(gradient descent (γ = 2), stochastic gradient descent (γ t = 0.01 ))
t

sgd
Andreas Eitel Uni FR DL 2017 (27)
 Neural Network backward pass

→ Now how do we compute the gradient for a network ?

l(f(x), y)

I Use the chain rule: smax smax

∂f (g(x)) ∂f (g(x)) ∂g(x)
f(x) a a
= (3)
∂x ∂g(x) ∂x W 1

I first compute loss on output layer

(2)
h (x) a
... a
... a
(2)
W
I then backpropagate to get
∂l(f (x), y) ∂l(f (x), y) 1
∂W(3)
and
∂a(3)
(1)
h (x) a
... a
... ai
(1)
W 1

x ... x ... x
1 j N

Andreas Eitel Uni FR DL 2017 (28)

 Neural Network backward pass

→ Now how do we compute the gradient for a network ?

l(f(x), y)

I gradient wrt. layer 3 weights:

smax smax
∂l(f (x), y)
=
∂l(f (x), y) ∂a(3) f(x) a a
(3)
W
∂W (3) ∂a(3) W(3) 1
I assuming l is NLL and softmax outputs, (2)
h (x) a
... a
... a
gradient wrt. layer 3 activation is:
(2)
∂l(f (x), y)
= −(y − f (x)) W 1
∂a(3)
y is one-hot encoded
(1)
h (x) a
... a
... ai
(1)
I gradient of a wrt. W(3) : W 1
∂a(3)
∂W(3)
= h(2) (x)T x ... x ... x
1 j N

→ recall
a(3) = W(3) h(2) (x) + b(3)

Andreas Eitel Uni FR DL 2017 (29)

 Neural Network backward pass

→ Now how do we compute the gradient for a network ?

l(f(x), y)
I gradient wrt. layer 3 weights:
∂l(f (x), y) ∂l(f (x), y) ∂a(3)
= smax smax
∂W (3) ∂a(3) W(3) f(x) a a
(3)
assuming l is NLL and softmax outputs,
W
I
1
gradient wrt. layer 3 activation is:
∂l(f (x), y)
(2)
h (x) a
... a
... a
= −(y − f (x)) (2)
∂a(3) W 1
I gradient of a wrt. W(3) : (1)
h (x) a
... a
... ai
∂a(3)
= h(2) (x)T (1)
∂W(3) W 1
I combined:
∂l(f (x), y) x ... x ... x
1 j N

= −(y − f (x))(h(2) (x))T

∂W(3)
→ recall
a(3) = W(3) h(2) (x) + b(3)

Andreas Eitel Uni FR DL 2017 (30)

 Neural Network backward pass

→ Now how do we compute the gradient for a network ?

I gradient wrt. layer 3 weights: l(f(x), y)

∂l(f (x), y) ∂l(f (x), y) ∂a(3)
=
∂W(3) ∂a(3) W(3)
I assuming l is NLL and softmax outputs, smax smax
gradient wrt. layer 3 activation is: f(x) a a
(3)
∂l(f (x), y)
= −(y − f (x)) W 1
∂a(3) (2)
h (x) a
... a
... a
I gradient of a wrt. W(3) : (2)
∂a(3) W 1

∂W(3)
= (h(2) (x))T (1)
h (x) a
... a
... ai
I gradient wrt. previous layer: (1)
W 1
∂l(f (x), y)
=
∂l(f (x), y) ∂a(3) x ... x ... x
1 j N

∂h(2) (x) ∂a(3) ∂h(2) (x)

 T ∂l(f (x), y)
= W(3) → recall
∂a(3)
a(3) = W(3) h(2) (x) + b(3)

Andreas Eitel Uni FR DL 2017 (31)

 Neural Network backward pass

→ Now how do we compute the gradient for a network ?

l(f(x), y)

I gradient wrt. layer 2 weights: smax smax

∂l(f (x), y) ∂l(f (x), y) ∂h(2) (x) ∂a(2) f(x) a a
= (3)
∂W (2) ∂h(2) (x) ∂a(2) W(2) W 1

I same schema as before just have to

(2)
h (x) a
... a
... a
consider computing derivative of activation (2)
∂h(2) (x) W 1
function
∂a(2)
, e.g. for sigmoid σ(·) (1)
h (x) a
... a
... ai
∂h(2) (x) (1)

∂a(2)
= σ(ai )(1 − ai ) W 1

I and backprop even further x ... x ... x

1 j N

→ recall
a(3) = W(3) h(2) (x) + b(3)

Andreas Eitel Uni FR DL 2017 (32)

 Gradient Checking

→ Backward-pass is just repeated application of the chain rule

→ However, there is a huge potential for bugs ...
→ Gradient checking to the rescue (simply check code via finite-differences):

Gradient Checking:
θ = (W(1) , b(1) , . . . ) init randomly
x ← init randomly ; y ← init randomly
∂l(fθ (x), y)
ganalytic ← compute gradient via backprop
∂θ
for i in #θ
I θ̂ = θ
I θ̂i = θ̂i + 
l(fθ̂ (x), y) − l(fθ (x), y)
I gnumeric =

I assert(kgnumeric − ganalytic k < )
I can also be used to test partial implementations
(i.e. layers, activation functions)
→ simply remove loss computation and backprop ones

Andreas Eitel Uni FR DL 2017 (33)

 Overfitting

I If you train the parameters of a large network θ = (W(1) , b(1) , . . . ) you

will see overfitting!
→ L(fθ (x), D)  L(fθ (x), V )
I This can be at least partly conquered with regularization

Andreas Eitel Uni FR DL 2017 (34)

 Overfitting

I If you train the parameters of a large network θ = (W(1) , b(1) , . . . ) you

will see overfitting!
→ L(fθ (x), D)  L(fθ (x), V )
I This can be at least partly conquered with regularization
I weight decay: change cost (and gradient)
N #θ
1 X 1 X
L(fθ , D) = min l(fθ (xi ), yi ) + kθi k2
N θ i=1 #θ i

→ enforces small weights (occams razor)

Andreas Eitel Uni FR DL 2017 (34)

 Overfitting

I If you train the parameters of a large network θ = (W(1) , b(1) , . . . ) you

will see overfitting!
→ L(fθ (x), D)  L(fθ (x), V )
I This can be at least partly conquered with regularization
I weight decay: change cost (and gradient)
N #θ
1 X 1 X
L(fθ , D) = min l(fθ (xi ), yi ) + kθi k2
N θ i=1 #θ i

→ enforces small weights (occams razor)

I dropout: kill ≈ 50% of the activations in each hidden layer during training
1
forward pass. Multiply hidden activations by during testing
2
→ prevents co-adaptation / enforces robustness to noise

Andreas Eitel Uni FR DL 2017 (34)

 Overfitting

I If you train the parameters of a large network θ = (W(1) , b(1) , . . . ) you

will see overfitting!
→ L(fθ (x), D)  L(fθ (x), V )
I This can be at least partly conquered with regularization
I weight decay: change cost (and gradient)
N #θ
1 X 1 X
L(fθ , D) = min l(fθ (xi ), yi ) + kθi k2
N θ i=1 #θ i

→ enforces small weights (occams razor)

I dropout: kill ≈ 50% of the activations in each hidden layer during training
1
forward pass. Multiply hidden activations by during testing
2
→ prevents co-adaptation / enforces robustness to noise
I Many, many more !

Andreas Eitel Uni FR DL 2017 (34)

 Assignment
I Implementation: Implement a simple feed-forward neural network by
completing the provided stub this includes:
I possibility to use 2-4 layers
I sigmoid/tanh and ReLU for the hidden layer
I softmax output layer
I optimization via gradient descent (gd)
I optimization via stochastic gradient descent (sgd)
I gradient checking code (!!!)
I weight initialization with random noise (!!!) (use normal distribution with
changing std. deviation for now)
I Bonus points for testing some advanced ideas:
I implement dropout, l2 regularization
I implement a different optimization scheme (rprop, rmsprop, adagrad)
I Code stub: https://github.com/aisrobots/dl_lab_2017
I Evaluation:
I Find good parameters (learning rate, number of iterations etc.) using a
validation set (usually take the last 10k examples from the training set)
I After optimizing parameters run on the full dataset and test once on the
test-set (you should be able to reach ≈ 1.6 − 1.8% error )
I Submission: Clone our github repo and send us the link to your
github/bitbucket repo including your solution code and report. Email to
[email protected], [email protected] and
[email protected]
Andreas Eitel Uni FR DL 2017 (35)
 Slide Information

I These slides have been created by Tobias Springenberg for the DL

Lab Course WS 2016/2017.

Andreas Eitel Uni FR DL 2017 (36)