A HANDS ON GUIDE TO

DEEP LEARNING

Andrew Beam, PhD

Instructor

Department of Biomedical Informatics

Harvard Medical School

BMI 703 - October 10th, 2018

twitter: @AndrewLBeam

WHY DEEP LEARNING?

DEEP LEARNING HAS MASTERED GO

Source: https://deepmind.com/blog/alphago-zero-learning-scratch/

Human data no longer needed

DEEP LEARNING HAS MASTERED GO

DEEP LEARNING CAN PLAY VIDEO GAMES

DEEP LEARNING CAN DIAGNOSE PATIENTS

Has the potential to change many medical specialities

DEEP LEARNING CAN DIAGNOSE PATIENTS

Has the potential to change many medical specialities

DEEP LEARNING CAN DIAGNOSE PATIENTS

What does this patient have?

A six-year old boy has a high fever that has lasted for three days. He has extremely red eyes and a rash on the main part of his body in addition to a swollen and red strawberry tongue. Remaining symptoms include swollen lymph nodes in the neck and Irritability

DEEP LEARNING CAN DIAGNOSE PATIENTS

What does this patient have?

Image credit: http://colah.github.io/posts/2015-08-Understanding-LSTMs/

DEEP LEARNING CAN DIAGNOSE PATIENTS

What does this patient have?

TRANSFORMING COMPUTATIONAL BIOLOGY

Not just medicine, but genomics too

More here: https://github.com/gokceneraslan/awesome-deepbio

EVERYONE IS USING DEEP LEARNING

WHAT IS A NEURAL NET?

MULTILAYER PERCEPTRONS (MLPs)

Most generic form of a neural net is the "multilayer perceptron"
Input undergoes a series of nonlinear transformation before a final classification layer

Image credit: http://cs231n.github.io

NEURAL NETWORKS ARE AN OLD IDEA

One of the very first ideas in machine learning and artificial intelligence

Date back to 1940s
Many cycles of boom and bust
Repeated promises of "true AI" that were unfulfilled followed by "AI winters"

Are today's neural nets any different than their predecessors?

"[The perceptron is] the embryo of an electronic computer that [the Navy] expects will be able to walk, talk, see, write, reproduce itself and be conscious of its existence." - Frank Rosenblatt, 1958

MILESTONES

IN THE BEGINNING... (1940s-1960s)

Rosenblatt's Perceptron, 1957

Initially very promising
Came with provably correct learning algorithm
Could recognize letters and numbers

THE FIRST AI WINTER (1969)

Minsky and Papert show that the perceptron can't even solve the XOR problem

Kills research on neural nets for the next 15-20 years

THE BACKPROPAGANDISTS EMERGE (1986)

Rumelhart, Hinton, and Willams show us how to train multilayered neural networks

THE PERCEPTRON AND BACKPROP

Pretend we just have two variables:

x_1,x_2

... and a class label:

y \in \{0,1\}

f(x_1,x_2) := w_0 + x_{i1} * w_1 + x_{i2} * w_2

Pretend we just have two variables:

x_1,x_2

... and a class label:

y \in \{0,1\}

... we construct a function to predict y from x:

THE PERCEPTRON AND BACKPROP

f(x_1,x_2) := w_0 + x_{i1} * w_1 + x_{i2} * w_2

Pretend we just have two variables:

x_1,x_2

... and a class label:

y \in \{0,1\}

... we construct a function to predict y from x:

p(x_1,x_2) := \frac{1}{1+\exp(-f(x_1,x_2))}

... and turn this into a probability using the logistic function:

THE PERCEPTRON AND BACKPROP

f(x_1,x_2) := w_0 + x_{i1} * w_1 + x_{i2} * w_2

Pretend we just have two variables:

x_1,x_2

... and a class label:

y \in \{0,1\}

... we construct a function to predict y from x:

p(x_1,x_2) := \frac{1}{1+\exp(-f(x_1,x_2))}

... and turn this into a probability using the logistic function:

\ell(w|x) := -y*\log(p(x)) - (1-y)*\log(1-p(x))

... and use Bernoulli negative log-likelihood as loss:

THE PERCEPTRON AND BACKPROP

f(x_1,x_2) := w_0 + x_{i1} * w_1 + x_{i2} * w_2

Pretend we just have two variables:

x_1,x_2

... and a class label:

y \in \{0,1\}

... we construct a function to predict y from x:

p(x_1,x_2) := \frac{1}{1+\exp(-f(x_1,x_2))}

... and turn this into a probability using the logistic function:

This is good old-fashioned logistic regression

\ell(w|x) := -y*\log(p(x)) - (1-y)*\log(1-p(x))

... and use Bernoulli negative log-likelihood as loss:

THE PERCEPTRON AND BACKPROP

How do we learn the "best" values for ?

\{w_0,w_1,w_2\}

THE PERCEPTRON AND BACKPROP

How do we learn the "best" values for ?

\{w_0,w_1,w_2\}

Gradient Decscent

Give weights random initial values
Evaluate partial derivative of each weight with respect negative log-likelihood at current weight value
Take a step in direction opposite to the gradient
Rinse and repeat

THE PERCEPTRON AND BACKPROP

How do we learn the "best" values for ?

\{w_0,w_1,w_2\}

Backprop is how we compute the gradients for our model's parameters. Later we will see why this is important

Gradient Decscent

Give weights random initial values
Evaluate partial derivative of each weight with respect negative log-likelihood at current weight value
Take a step in direction opposite to the gradient
Rinse and repeat

THE PERCEPTRON AND BACKPROP

NEURAL NET EXERCISE

MLPs AND MODERN NEURAL NETS

LOGISTIC REGRESSION -> MLP

With a small change, we can turn our logistic regression model into a neural net

Instead of just one linear combination, we are going to take several, each with a different set of weights (called a hidden unit)
Each linear combination will be followed by a nonlinear activation
Each of these nonlinear features will be fed into the logistic regression classifier
All of the weights are learned end-to-end via SGD

MLPs learn a set of nonlinear features directly from data

"Feature learning" is the hallmark of deep learning approachs

MLPs

Increase the depth to increase network capacity

LINEAR ALGEBRA & NEURAL NETS

Essentially all neural nets can be described in terms of a series of matrix multiplications

H_1 = h_1(X) = \phi(X * W_{1})

Input to hidden layer, with 100 hidden units:

where:

- is a matrix of weights

- is matrix containing the activations for all 100 units for each sample and

- is an element-wise activation function

W_1

\phi()

H_1

LINEAR ALGEBRA & NEURAL NETS

Essentially all neural nets can be described in terms of a series of matrix multiplications

H_1 = h_1(X) = \phi(X * W_{1})

Input to hidden layer, with 100 hidden units:

H_2 = h_2(X) = \phi(H_1 * W_{1})

Hidden to hidden layer:

MODERN DEEP LEARNING

Several key advancements have enabled the modern deep learning revolution

Advent of massively parallel computing by GPUs. Enabled training of huge neural nets on extremely large datasets

Several key advancements have enabled the modern deep learning revolution

Methodological advancements have made deeper networks easier to train

Architecture

Optimizers

Activation Functions

MODERN DEEP LEARNING

Several key advancements have enabled the modern deep learning revolution

Robust frameworks and abstractions make iteration faster and less error prone

Automatic differentiation allows easy prototyping

MODERN DEEP LEARNING

CONVOLUTIONAL NEURAL NETWORKS

CONVOLUTIONAL NEURAL NETS (CNNs)

Dates back to the late 1980s

Invented by in 1989 Yann Lecun at Bell Labs - "Lenet"
Integrated into handwriting recognition systems
in the 90s
Huge flurry of activity after the Alexnet paper

THE BREAKTHROUGH (2012)

Imagenet Database

Millions of labeled images
Objects in images fall into 1 of a possible 1,000 categories
Relatively high-resolution
Bounding boxes giving exact location of object - useful for both classification and localization

Large Scale Visual

Recognition Challenge (ILSVRC)

Annual Imagenet Challenge starting in 2010
Successor to smaller PASCAL VOC challenge
Many tracks including classification and localization
Standardized training and test set. Competitors upload predictions for test set and are automatically scored

THE BREAKTHROUGH (2012)

Pivotal event occurred in an image recognition contest which brought together 3 critical ingredients for the first time:

Massive amounts of labeled images
Training with GPUs
Methodological innovations that enabled training deeper networks while minimizing overfitting

THE BREAKTHROUGH (2012)

In 2011, a misclassification rate of 25% was near state of the art on ILSVRC

In 2012, Geoff Hinton and two graduate students, Alex Krizhevsky and Ilya Sutskever, entered ILSVRC with one of the first deep neural networks trained on GPUs, now known as "Alexnet"

Result: An error rate of 16%, nearly half what the second place entry was able to achieve.

The computer vision world immediately took notice

THE ILSVRC AFTERMATH (2012-2014)

Alexnet paper has ~ 16,000 citations since being published in 2012!

Q: What was the major flaw with our MLP dog vs. cat classifier?

CONVOLUTIONAL NEURAL NETS (CNNs)

Q: What was the major flaw with our MLP dog vs. cat classifier?

CONVOLUTIONAL NEURAL NETS (CNNs)

What is this a picture of?

Q: What was the major flaw with our MLP dog vs. cat classifier?

CONVOLUTIONAL NEURAL NETS (CNNs)

What is this a picture of?

Q: What was the major flaw with our MLP dog vs. cat classifier?

CONVOLUTIONAL NEURAL NETS (CNNs)

What is this a picture of?

MLPs throw away too much information!

CONVOLUTIONAL NEURAL NETS

Images are just 2D arrays of numbers

Goal is to build f(image) = 1

CONVOLUTIONAL NEURAL NETS

CNNs look at small connected groups of pixels using "filters"

Image credit: http://deeplearning.stanford.edu/wiki/index.php/Feature_extraction_using_convolution

Images have a local correlation structure

Near by pixels are likely to be more similar than pixels that are far away

CNNs exploit this through convolutions of small image patches

CONVOLUTIONAL NEURAL NETS

Example convolution

CONVOLUTIONAL NEURAL NETS

Convolution + pooling + activation = CNN

Image credit: http://cs231n.github.io/convolutional-networks/

CONVOLUTIONAL NEURAL NETS

Pooling provides spatial invariance

Image credit: http://cs231n.github.io/convolutional-networks/

CONVOLUTIONAL NEURAL NETS (CNNs)

CNN formula is relatively simple

Image credit: http://cs231n.github.io/convolutional-networks/

CONVOLUTIONAL NEURAL NETS

Data augmentation mimics the image generative process

Image credit: http://slideplayer.com/slide/8370683/

Drastically "expands" training set size
Improves generalization
Works if it doesn't "break" image -> label relationship

CNN EXERCISE

CNNS AREN'T MAGIC

Based on solid image priors
Learns a hierarchical set of filters
Exploit properties of images, e.g. local correlations and invariances
Mimic generative distribution with augmentation to reduce over fitting
Results in end-to-end visual recognition system trained with SGD on GPUs: pixels in -> classifications out

CNNs exploit strong prior information about images

SUMMARY & CONCLUSIONS

DEEP LEARNING AND YOU

Barrier to entry for deep learning is actually low

... but a few things might stand in your way:

Need to make sure your problem is a good fit
- Lots of labeled data and appropriate signal/noise ratio
Access to GPUs
Must "speak the language"
- Many design choices and hyper parameter selections
Know how to "babysit" the model during learning phase

DEEP LEARNING AND YOU

Barrier to entry for deep learning is actually low

... but a few things might stand in your way:

Need to make sure your problem is a good fit
- Lots of labeled data and appropriate signal/noise ratio
Access to GPUs
Must "speak the language"
- Many design choices and hyper parameter selections
Know how to "babysit" the model during learning phase

HOW CAN YOU STAY CURRENT?

The field moves fast, staying up to date can be challenging

http://beamandrew.github.io/deeplearning/2017/02/23/deep_learning_101_part1.html

CONCLUSIONS

Could potentially impact many fields -> understand concepts so you have deep learning "insurance"
Long history and connections to other models and fields
Prereqs: Data (lots) + GPUs (more = better)
Deep learning models are like legos, but you need to know what blocks you have and how they fit together
Need to have a sense of sensible default parameter values to get started
"Babysitting" the learning process is a skill