2014 Lasker Awards:

2014 Lasker~DeBakey Clinical Medical Research Award

Deep brain stimulation for Parkinson’s disease

Alim Louis Benabid

Joseph Fourier University, Grenoble

Mahlon R. DeLong
Emory University School of Medicine
The 2014 Lasker~DeBakey Clinical Medical Research Award honors two scientists who developed deep
brain stimulation of the subthalamic nucleus, a surgical technique that reduces tremors and restores motor
function in patients with advanced Parkinson's disease. Mahlon R. DeLong (Emory University School of
Medicine) formulated a new model for the brain's circuitry and exposed a fresh target for this illness.
Alim Louis Benabid (Joseph Fourier University, Grenoble) devised an effective and reversible
intervention that remedies neuronal misfirings. Their work has culminated in an effective treatment for
more than 100,000 individuals worldwide with severe illness who suffer from complications of levodopa
therapy.
Parkinson's disease (PD), perhaps best known for its tremor, slows and stiffens movements. From the
1940s through the 1960s, surgeons battled the ailment by destroying regions of the brain, chosen more by
trial and error than by a clear understanding of neural misbehavior. The so-called lesions created by these
operations often delivered spectacular and stable effects, counteracting the tremor and, to some extent,
other features of PD. Even slight misplacement, however, brought complications rather than benefits.
Such damage was permanent, as dead tissue could not be revived.
Award presentation by Greg Petsko

In 1972, the novelist Michael Crichton, whose Harvard Medical School education seems primarily to
have turned him into a techno-Luddite, wrote a novel called The Terminal Man. It told the story of Harry
Benson, who has electrodes implanted into his brain to suppress his psychomotor epileptic seizures. The
electrodes are controlled by a small computer with a power pack under the skin in his shoulder.
Unfortunately, this being a Michael Crichton novel, the technology is flawed and Benson is psychotic,
and — well, you can probably guess how this turns out.
That same year, 1972, a young medical doctor in France named Alim-Louis Benabid became a staff
neurosurgeon at Joseph Fourier University in Grenoble while also beginning his studies for an eventual
doctorate in physics. And also in 1972, in the United States, another young doctor, Mahlon DeLong, was
finishing up his time as a research associate in the National Institute of Mental Health.
Eventually, the work of these two men would coalesce into something strikingly similar, sort of, to the
device in The Terminal Man. But its use, Michael Crichton could never have dreamed up: a new therapy
for Parkinson's disease.
Alim Louis Benabid
Acceptance remarks, 2014 Lasker Awards Ceremony

Nature Medicine Interview

I was born in France during World War II and raised as a teenager during the Algerian War. My brothers
and I were told by my father and my mother that peace was the highest value to pursue. My father, who
was an MD, wanted me to be a physician and my mother who was a nurse wanted me to be a physicist.
To preserve peace, I had no other choice than being both.
I would say, then, that my life has been a science fiction novel, my story being based on the patient's
problems and expectations, the use of increasingly sophisticated tools in the battles, the thrill of facing
challenges, and the delight of solving them, with sometimes happy ends for the patients, and sometimes
not.
Mahlon R. DeLong
Acceptance remarks, 2014 Lasker Awards Ceremony

Nature Medicine Interview

I don't recall when I realized I wanted to do research, but I have always enjoyed understanding how things
work. A growing fascination with how the brain controls behavior led me to medicine and then to
neurology. This took a clear direction when I found a choice research position at the NIH in the
laboratory of renowned researcher Edward Evarts. Because the other obvious brain regions were already
assigned to other fellows, I was asked to work on the basal ganglia, a cluster of poorly understood brain
structures, and to determine their role in the control of bodily movements.
For years I never thought about finding a cure for anything but was fascinated by the progress in
understanding the anatomical relations and functions of these mysterious structures. Almost two decades
later, a new animal model of Parkinson's was developed, made possible by the discovery of a toxin that
produced clinical features and pathology closely resembling Parkinson's, resulting from a loss of
dopamine in the basal ganglia. Our basic research provided a solid foundation for what was to come. We
found abnormal neuronal activity throughout the motor circuit of the basal ganglia and 'struck gold' by
making a lesion in the subthalamic nucleus, which, by interrupting the circuit, almost immediately
restored movement and eliminated tremor and rigidity. Most importantly, the findings provided insight
into the neuronal dysfunction responsible for Parkinson's and a strong rational for surgical intervention as
well as a novel target.
Having spent much of my early career in both basic research and clinical work, I have been fortunate to
see how the basic science contributes to patient care. I am also aware of how far this field, that we now
call neuromodulation, has progressed.
Lasker Awards honor individuals who contribute critically to a research problem — but they also
highlight the larger progress in a given field, in our case, using targeted electrical stimulation to restore
function in the disrupted neural networks responsible for both neurologic and psychiatric disorders.
For nearly a century, neurologists learned about brain function by correlating the behavioral changes
caused by strokes with the damage to specific regions of the brain. With deep brain stimulation, we are
now not only able to treat disorders such as Parkinson's — we're also able to learn about brain networks
and their function by observing the effects of targeted stimulation and activation. Whereas we previously
thought about brain 'centers', we now focus on specific brain circuits and how they 'go wrong' in
neurologic and psychiatric disorders. Understanding how neural circuits are disrupted by disease and how
lesions and electrical stimulation can improve function is still poorly understood. Understanding brain
function and its disorders is the last and most challenging frontier of medicine.
Some of our real progress and 'breakthroughs', are made, I believe, serendipitously, while we are doing
other things — like just trying to understand how things work. I sometimes wonder whether, in our
urgency to find cures, we sometimes omit the basic science necessary to understand how things work —
and by doing so, may miss the chance of finding a new treatment.
Key publications of Mahlon R. DeLong
Alexander, G.E., DeLong, M.R., and Strick, P.L. (1986). Parallel organization of functionally segregated
circuits linking basal ganglia and cortex. Ann. Rev. Neurosci. 9, 357-381.
Miller, W.C. and DeLong, M.R. (1987). Altered tonic activity of neurons in the globus pallidus and
subthalamic nucleus in the primate MPTP model of parkinsonism. In: Advances of Behavioral Biology.
M.B. Carpenter and A. Jayaraman (eds.), Plenum Publishing Corp., 32, 415-427, New York.
DeLong, M.R. (1990). Primate models of movement disorders of basal ganglia origin. Trends Neurosci.
13, 281-285.
Bergman, H., Wichmann, T. and DeLong, M.R. (1990). Reversal of experimental parkinsonism by
lesions of the subthalamic nucleus. Science. 249, 1436-1438.
Bergman, H.G., Wichmann, T., and DeLong, M.R. (1994). The primate subthalamic nucleus: II. Neural
activity in the subthalamic nucleus and pallidum in the MPTP model of parkinsonism. J. Neurophysiol.
72, 507-520.
DeLong, M.R. and Wichmann, T. (2007). Circuits and circuit disorders of the basal ganglia. Arch.
Neurol. 64, 20-24.
Mahlon R. DeLong
Acceptance remarks, 2014 Lasker Awards Ceremony

Nature Medicine Interview

I don't recall when I realized I wanted to do research, but I have always enjoyed understanding how things
work. A growing fascination with how the brain controls behavior led me to medicine and then to
neurology. This took a clear direction when I found a choice research position at the NIH in the
laboratory of renowned researcher Edward Evarts. Because the other obvious brain regions were already
assigned to other fellows, I was asked to work on the basal ganglia, a cluster of poorly understood brain
structures, and to determine their role in the control of bodily movements.
For years I never thought about finding a cure for anything but was fascinated by the progress in
understanding the anatomical relations and functions of these mysterious structures. Almost two decades
later, a new animal model of Parkinson's was developed, made possible by the discovery of a toxin that
produced clinical features and pathology closely resembling Parkinson's, resulting from a loss of
dopamine in the basal ganglia. Our basic research provided a solid foundation for what was to come. We
found abnormal neuronal activity throughout the motor circuit of the basal ganglia and 'struck gold' by
making a lesion in the subthalamic nucleus, which, by interrupting the circuit, almost immediately
restored movement and eliminated tremor and rigidity. Most importantly, the findings provided insight
into the neuronal dysfunction responsible for Parkinson's and a strong rational for surgical intervention as
well as a novel target.
Having spent much of my early career in both basic research and clinical work, I have been fortunate to
see how the basic science contributes to patient care. I am also aware of how far this field, that we now
call neuromodulation, has progressed.
Lasker Awards honor individuals who contribute critically to a research problem — but they also
highlight the larger progress in a given field, in our case, using targeted electrical stimulation to restore
function in the disrupted neural networks responsible for both neurologic and psychiatric disorders.
For nearly a century, neurologists learned about brain function by correlating the behavioral changes
caused by strokes with the damage to specific regions of the brain. With deep brain stimulation, we are
now not only able to treat disorders such as Parkinson's — we're also able to learn about brain networks
and their function by observing the effects of targeted stimulation and activation. Whereas we previously
thought about brain 'centers', we now focus on specific brain circuits and how they 'go wrong' in
neurologic and psychiatric disorders. Understanding how neural circuits are disrupted by disease and how
lesions and electrical stimulation can improve function is still poorly understood. Understanding brain
function and its disorders is the last and most challenging frontier of medicine.
Some of our real progress and 'breakthroughs', are made, I believe, serendipitously, while we are doing
other things — like just trying to understand how things work. I sometimes wonder whether, in our
urgency to find cures, we sometimes omit the basic science necessary to understand how things work —
and by doing so, may miss the chance of finding a new treatment.
Key publications of Mahlon R. DeLong
Alexander, G.E., DeLong, M.R., and Strick, P.L. (1986). Parallel organization of functionally segregated
circuits linking basal ganglia and cortex. Ann. Rev. Neurosci. 9, 357-381.
Miller, W.C. and DeLong, M.R. (1987). Altered tonic activity of neurons in the globus pallidus and
subthalamic nucleus in the primate MPTP model of parkinsonism. In: Advances of Behavioral Biology.
M.B. Carpenter and A. Jayaraman (eds.), Plenum Publishing Corp., 32, 415-427, New York.
DeLong, M.R. (1990). Primate models of movement disorders of basal ganglia origin. Trends Neurosci.
13, 281-285.
Bergman, H., Wichmann, T. and DeLong, M.R. (1990). Reversal of experimental parkinsonism by
lesions of the subthalamic nucleus. Science. 249, 1436-1438.
Bergman, H.G., Wichmann, T., and DeLong, M.R. (1994). The primate subthalamic nucleus: II. Neural
activity in the subthalamic nucleus and pallidum in the MPTP model of parkinsonism. J. Neurophysiol.
72, 507-520.
DeLong, M.R. and Wichmann, T. (2007). Circuits and circuit disorders of the basal ganglia. Arch.
Neurol. 64, 20-24.
Deep Learning
What Is Deep Learning?
3 things you need to know

Deep learning is a machine learning technique that teaches computers to do what comes naturally to
humans: learn by example. Deep learning is a key technology behind driverless cars, enabling them to
recognize a stop sign, or to distinguish a pedestrian from a lamppost. It is the key to voice control in
consumer devices like phones, tablets, TVs, and hands-free speakers. Deep learning is getting lots of
attention lately and for good reason. It’s achieving results that were not possible before.
In deep learning, a computer model learns to perform classification tasks directly from images, text, or
sound. Deep learning models can achieve state-of-the-art accuracy, sometimes exceeding human-level
performance. Models are trained by using a large set of labeled data and neural network architectures that
contain many layers.
 Why Deep Learning Matters
How does deep learning attain such impressive results?
In a word, accuracy. Deep learning achieves recognition accuracy at higher levels than ever before. This
helps consumer electronics meet user expectations, and it is crucial for safety-critical applications like
driverless cars. Recent advances in deep learning have improved to the point where deep learning
outperforms humans in some tasks like classifying objects in images.
While deep learning was first theorized in the 1980s, there are two main reasons it has only recently
become useful:
1. Deep learning requires large amounts of labeled data. For example, driverless car
development requires millions of images and thousands of hours of video.
2. Deep learning requires substantial computing power. High-performance GPUs have a parallel
architecture that is efficient for deep learning. When combined with clusters or cloud
computing, this enables development teams to reduce training time for a deep learning
network from weeks to hours or less.

Deep Learning with MATLAB (ebook)

Examples of Deep Learning at Work
Deep learning applications are used in industries from automated driving to medical devices.
Automated Driving: Automotive researchers are using deep learning to automatically detect objects such
as stop signs and traffic lights. In addition, deep learning is used to detect pedestrians, which helps
decrease accidents.
Aerospace and Defense: Deep learning is used to identify objects from satellites that locate areas of
interest, and identify safe or unsafe zones for troops.
Medical Research: Cancer researchers are using deep learning to automatically detect cancer cells. Teams
at UCLA built an advanced microscope that yields a high-dimensional data set used to train a deep
learning application to accurately identify cancer cells.
Industrial Automation: Deep learning is helping to improve worker safety around heavy machinery by
automatically detecting when people or objects are within an unsafe distance of machines.
Electronics: Deep learning is being used in automated hearing and speech translation. For example, home
assistance devices that respond to your voice and know your preferences are powered by deep learning
applications.
How Deep Learning Works
Most deep learning methods use neural network architectures, which is why deep learning models are
often referred to as deep neural networks.
The term “deep” usually refers to the number of hidden layers in the neural network. Traditional neural
networks only contain 2-3 hidden layers, while deep networks can have as many as 150.
Deep learning models are trained by using large sets of labeled data and neural network architectures that
learn features directly from the data without the need for manual feature extraction.
Figure 1: Neural networks, which are organized in layers consisting of a set of interconnected nodes.
Networks can have tens or hundreds of hidden layers.
One of the most popular types of deep neural networks is known as convolutional neural
networks (CNN or ConvNet). A CNN convolves learned features with input data, and uses 2D
convolutional layers, making this architecture well suited to processing 2D data, such as images.
CNNs eliminate the need for manual feature extraction, so you do not need to identify features used to
classify images. The CNN works by extracting features directly from images. The relevant features are
not pretrained; they are learned while the network trains on a collection of images. This automated feature
extraction makes deep learning models highly accurate for computer vision tasks such as object
classification.

Figure 2: Example of a network with many convolutional layers. Filters are applied to each training image
at different resolutions, and the output of each convolved image serves as the input to the next layer.
CNNs learn to detect different features of an image using tens or hundreds of hidden layers. Every hidden
layer increases the complexity of the learned image features. For example, the first hidden layer could
learn how to detect edges, and the last learns how to detect more complex shapes specifically catered to
the shape of the object we are trying to recognize.
What's the Difference Between Machine Learning and Deep Learning?
Deep learning is a specialized form of machine learning. A machine learning workflow starts with
relevant features being manually extracted from images. The features are then used to create a model that
categorizes the objects in the image. With a deep learning workflow, relevant features are automatically
extracted from images. In addition, deep learning performs “end-to-end learning” – where a network is
given raw data and a task to perform, such as classification, and it learns how to do this automatically.
Another key difference is deep learning algorithms scale with data, whereas shallow learning converges.
Shallow learning refers to machine learning methods that plateau at a certain level of performance when
you add more examples and training data to the network.
A key advantage of deep learning networks is that they often continue to improve as the size of your data
increases.

Figure 3. Comparing a machine learning approach to categorizing vehicles (left) with deep learning
(right).
In machine learning, you manually choose features and a classifier to sort images. With deep learning,
feature extraction and modeling steps are automatic.
Object Recognition: Deep Learning and Machine Learning for Computer Vision (26:57)
Introduction to Deep Learning: What Are Convolutional Neural Networks? (4:44)
Choosing Between Machine Learning and Deep Learning
Machine learning offers a variety of techniques and models you can choose based on your application, the
size of data you're processing, and the type of problem you want to solve. A successful deep learning
application requires a very large amount of data (thousands of images) to train the model, as well
as GPUs, or graphics processing units, to rapidly process your data.
When choosing between machine learning and deep learning, consider whether you have a high-
performance GPU and lots of labeled data. If you don’t have either of those things, it may make more
sense to use machine learning instead of deep learning. Deep learning is generally more complex, so
you’ll need at least a few thousand images to get reliable results. Having a high-performance GPU means
the model will take less time to analyze all those images.
How to Create and Train Deep Learning Models
The three most common ways people use deep learning to perform object classification are:
Training from Scratch
To train a deep network from scratch, you gather a very large labeled data set and design a network
architecture that will learn the features and model. This is good for new applications, or applications that
will have a large number of output categories. This is a less common approach because with the large
amount of data and rate of learning, these networks typically take days or weeks to train.
Deep Learning with MATLAB: Deep Learning in 11 Lines of MATLAB Code (2:38)
Transfer Learning
Most deep learning applications use the transfer learning approach, a process that involves fine-tuning a
pretrained model. You start with an existing network, such as AlexNet or GoogLeNet, and feed in new
data containing previously unknown classes. After making some tweaks to the network, you can now
perform a new task, such as categorizing only dogs or cats instead of 1000 different objects. This also has
the advantage of needing much less data (processing thousands of images, rather than millions), so
computation time drops to minutes or hours.
Transfer learning requires an interface to the internals of the pre-existing network, so it can be surgically
modified and enhanced for the new task. MATLAB® has tools and functions designed to help you do
transfer learning.
Deep Learning with MATLAB: Transfer Learning in 10 Lines of MATLAB Code (3:59)
Feature Extraction
A slightly less common, more specialized approach to deep learning is to use the network as a feature
extractor. Since all the layers are tasked with learning certain features from images, we can pull these
features out of the network at any time during the training process. These features can then be used as
input to a machine learning model such as support vector machines (SVM).
Accelerating Deep Learning Models with GPUs
Training a deep learning model can take a long time, from days to weeks. Using GPU acceleration can
speed up the process significantly. Using MATLAB with a GPU reduces the time required to train a
network and can cut the training time for an image classification problem from days down to hours. In
training deep learning models, MATLAB uses GPUs (when available) without requiring you to
understand how to program GPUs explicitly.

Figure 4. Neural Network Toolbox commands for training your own CNN from scratch or using a
pretrained model for transfer learning.
Deep Learning Applications
Pretrained deep neural network models can be used to quickly apply deep learning to your problems by
performing transfer learning or feature extraction. For MATLAB users, some available models include
AlexNet, VGG-16, and VGG-19, as well as Caffe models (for example, from Caffe Model Zoo) imported
using importCaffeNetwork.

Use AlexNet to Recognize Objects with Your Webcam

Use MATLAB, a simple webcam, and a deep neural network to identify objects in your surroundings.
Learn more

Example: Object Detection Using Deep Learning

In addition to object recognition, which identifies a specific object in an image or video, deep learning
can also be used for object detection. Object detection means recognizing and locating the object in a
scene, and it allows for multiple objects to be located within the image.
See example
Deep Learning with MATLAB
MATLAB makes deep learning easy. With tools and functions for managing large data sets, MATLAB
also offers specialized toolboxes for working with machine learning, neural networks, computer vision,
and automated driving.
With just a few lines of code, MATLAB let you do deep learning without being an expert. Get started
quickly, create and visualize models, and deploy models to servers and embedded devices.
Teams are successful using MATLAB for deep learning because it lets you:
1. Create and Visualize Models with Just a Few Lines of Code.
MATLAB lets you build deep learning models with minimal code. With MATLAB, you can
quickly import pretrained models and visualize and debug intermediate results as you adjust
training parameters.
2. Perform Deep Learning Without Being an Expert.
You can use MATLAB to learn and gain expertise in the area of deep learning. Most of us
have never taken a course in deep learning. We have to learn on the job. MATLAB makes
learning about this field practical and accessible. In addition, MATLAB enables domain
experts to do deep learning – instead of handing the task over to data scientists who may not
know your industry or application.
3. Automate Ground Truth Labeling of Images and Video.
MATLAB enables users to interactively label objects within images and can automate
ground truth labeling within videos for training and testing deep learning models. This
interactive and automated approach can lead to better results in less time.
4. Integrate Deep Learning in a Single Workflow.
MATLAB can unify multiple domains in a single workflow. With MATLAB, you can do
your thinking and programming in one environment. It offers tools and functions for deep
learning, and also for a range of domains that feed into deep learning algorithms, such as
signal processing, computer vision, and data analytics.
With MATLAB, you can integrate results into your existing applications. MATLAB automates deploying
your deep learning models on enterprise systems, clusters, clouds, and embedded devices.
See more deep learning features and get example code.
Related products: MATLAB, Computer Vision System Toolbox™, Statistics and Machine Learning
Toolbox™, Neural Network Toolbox™, and Automated Driving System Toolbox™.
 Learn More About Deep Learning

3:33

Introduction to Deep Learning: What Is Deep Learning?

Explore deep learning fundamentals in this MATLAB Tech Talk. You’ll learn why deep learning has
become so popular, and you’ll walk through 3 concepts: what deep learning is, how it is used in the real
world, and how you can get started.

2:38

Deep Learning with MATLAB: Deep Learning in 11 Lines of MATLAB Code

See how to use MATLAB, a simple webcam, and a deep neural network to identify objects in your
surroundings. This demo uses AlexNet, a pretrained deep convolutional neural network that has been
trained on over a million images.

3:59

Deep Learning with MATLAB: Transfer Learning in 10 Lines of MATLAB Code

Learn how to use transfer learning in MATLAB to re-train deep learning networks created by experts for
your own data or task.
Cancer Diagnostics with Deep Learning and Photonic Time Stretch
An imaging system combining flow cytometry, photonic time stretch, and machine learning algorithms
enables UCLA researchers to classify cancer cells in blood samples without using biomarker labels.

7:49

Image Recognition Using Machine Learning

Use machine learning techniques in MATLAB to recognize scenes based on unique features within
images.

Deep Learning for Computer Vision with MATLAB

Using a simple object detection and recognition example, this article illustrates how easy it is to use
MATLAB® for deep learning, even without extensive knowledge of advanced computer vision
algorithms or neural networks.

Why Deep Learning Matters

How does deep learning attain such impressive results?

In a word, accuracy. Deep learning achieves recognition accuracy at higher levels than ever before. This
helps consumer electronics meet user expectations, and it is crucial for safety-critical applications like
driverless cars. Recent advances in deep learning have improved to the point where deep learning
outperforms humans in some tasks like classifying objects in images.
While deep learning was first theorized in the 1980s, there are two main reasons it has only recently
become useful:

1. Deep learning requires large amounts of labeled data. For example, driverless car
development requires millions of images and thousands of hours of video.
2. Deep learning requires substantial computing power. High-performance GPUs have a parallel
architecture that is efficient for deep learning. When combined with clusters or cloud
computing, this enables development teams to reduce training time for a deep learning
network from weeks to hours or less.

Deep Learning with MATLAB (ebook)

Examples of Deep Learning at Work
Deep learning applications are used in industries from automated driving to medical devices.
Automated Driving: Automotive researchers are using deep learning to automatically detect objects such
as stop signs and traffic lights. In addition, deep learning is used to detect pedestrians, which helps
decrease accidents.
Aerospace and Defense: Deep learning is used to identify objects from satellites that locate areas of
interest, and identify safe or unsafe zones for troops.
Medical Research: Cancer researchers are using deep learning to automatically detect cancer cells. Teams
at UCLA built an advanced microscope that yields a high-dimensional data set used to train a deep
learning application to accurately identify cancer cells.
Industrial Automation: Deep learning is helping to improve worker safety around heavy machinery by
automatically detecting when people or objects are within an unsafe distance of machines.
Electronics: Deep learning is being used in automated hearing and speech translation. For example, home
assistance devices that respond to your voice and know your preferences are powered by deep learning
applications.
How Deep Learning Works
Most deep learning methods use neural network architectures, which is why deep learning models are
often referred to as deep neural networks.
The term “deep” usually refers to the number of hidden layers in the neural network. Traditional neural
networks only contain 2-3 hidden layers, while deep networks can have as many as 150.
Deep learning models are trained by using large sets of labeled data and neural network architectures that
learn features directly from the data without the need for manual feature extraction.
Figure 1: Neural networks, which are organized in layers consisting of a set of interconnected nodes.
Networks can have tens or hundreds of hidden layers.
One of the most popular types of deep neural networks is known as convolutional neural
networks (CNN or ConvNet). A CNN convolves learned features with input data, and uses 2D
convolutional layers, making this architecture well suited to processing 2D data, such as images.
CNNs eliminate the need for manual feature extraction, so you do not need to identify features used to
classify images. The CNN works by extracting features directly from images. The relevant features are
not pretrained; they are learned while the network trains on a collection of images. This automated feature
extraction makes deep learning models highly accurate for computer vision tasks such as object
classification.

Figure 2: Example of a network with many convolutional layers. Filters are applied to each training image
at different resolutions, and the output of each convolved image serves as the input to the next layer.
CNNs learn to detect different features of an image using tens or hundreds of hidden layers. Every hidden
layer increases the complexity of the learned image features. For example, the first hidden layer could
learn how to detect edges, and the last learns how to detect more complex shapes specifically catered to
the shape of the object we are trying to recognize.
What's the Difference Between Machine Learning and Deep Learning?
Deep learning is a specialized form of machine learning. A machine learning workflow starts with
relevant features being manually extracted from images. The features are then used to create a model that
categorizes the objects in the image. With a deep learning workflow, relevant features are automatically
extracted from images. In addition, deep learning performs “end-to-end learning” – where a network is
given raw data and a task to perform, such as classification, and it learns how to do this automatically.
Another key difference is deep learning algorithms scale with data, whereas shallow learning converges.
Shallow learning refers to machine learning methods that plateau at a certain level of performance when
you add more examples and training data to the network.
A key advantage of deep learning networks is that they often continue to improve as the size of your data
increases.

Figure 3. Comparing a machine learning approach to categorizing vehicles (left) with deep learning
(right).
In machine learning, you manually choose features and a classifier to sort images. With deep learning,
feature extraction and modeling steps are automatic.
Object Recognition: Deep Learning and Machine Learning for Computer Vision (26:57)
Introduction to Deep Learning: What Are Convolutional Neural Networks? (4:44)
Choosing Between Machine Learning and Deep Learning
Machine learning offers a variety of techniques and models you can choose based on your application, the
size of data you're processing, and the type of problem you want to solve. A successful deep learning
application requires a very large amount of data (thousands of images) to train the model, as well
as GPUs, or graphics processing units, to rapidly process your data.
When choosing between machine learning and deep learning, consider whether you have a high-
performance GPU and lots of labeled data. If you don’t have either of those things, it may make more
sense to use machine learning instead of deep learning. Deep learning is generally more complex, so
you’ll need at least a few thousand images to get reliable results. Having a high-performance GPU means
the model will take less time to analyze all those images.
How to Create and Train Deep Learning Models
The three most common ways people use deep learning to perform object classification are:
Training from Scratch
To train a deep network from scratch, you gather a very large labeled data set and design a network
architecture that will learn the features and model. This is good for new applications, or applications that
will have a large number of output categories. This is a less common approach because with the large
amount of data and rate of learning, these networks typically take days or weeks to train.
Deep Learning with MATLAB: Deep Learning in 11 Lines of MATLAB Code (2:38)
Transfer Learning
Most deep learning applications use the transfer learning approach, a process that involves fine-tuning a
pretrained model. You start with an existing network, such as AlexNet or GoogLeNet, and feed in new
data containing previously unknown classes. After making some tweaks to the network, you can now
perform a new task, such as categorizing only dogs or cats instead of 1000 different objects. This also has
the advantage of needing much less data (processing thousands of images, rather than millions), so
computation time drops to minutes or hours.
Transfer learning requires an interface to the internals of the pre-existing network, so it can be surgically
modified and enhanced for the new task. MATLAB® has tools and functions designed to help you do
transfer learning.
Deep Learning with MATLAB: Transfer Learning in 10 Lines of MATLAB Code (3:59)
Feature Extraction
A slightly less common, more specialized approach to deep learning is to use the network as a feature
extractor. Since all the layers are tasked with learning certain features from images, we can pull these
features out of the network at any time during the training process. These features can then be used as
input to a machine learning model such as support vector machines (SVM).
Accelerating Deep Learning Models with GPUs
Training a deep learning model can take a long time, from days to weeks. Using GPU acceleration can
speed up the process significantly. Using MATLAB with a GPU reduces the time required to train a
network and can cut the training time for an image classification problem from days down to hours. In
training deep learning models, MATLAB uses GPUs (when available) without requiring you to
understand how to program GPUs explicitly.

Figure 4. Neural Network Toolbox commands for training your own CNN from scratch or using a
pretrained model for transfer learning.
Deep Learning Applications
Pretrained deep neural network models can be used to quickly apply deep learning to your problems by
performing transfer learning or feature extraction. For MATLAB users, some available models include
AlexNet, VGG-16, and VGG-19, as well as Caffe models (for example, from Caffe Model Zoo) imported
using importCaffeNetwork.
Use AlexNet to Recognize Objects with Your Webcam
Use MATLAB, a simple webcam, and a deep neural network to identify objects in your surroundings.
Learn more

Example: Object Detection Using Deep Learning

In addition to object recognition, which identifies a specific object in an image or video, deep learning
can also be used for object detection. Object detection means recognizing and locating the object in a
scene, and it allows for multiple objects to be located within the image.
See example
Deep Learning with MATLAB
MATLAB makes deep learning easy. With tools and functions for managing large data sets, MATLAB
also offers specialized toolboxes for working with machine learning, neural networks, computer vision,
and automated driving.
With just a few lines of code, MATLAB let you do deep learning without being an expert. Get started
quickly, create and visualize models, and deploy models to servers and embedded devices.
Teams are successful using MATLAB for deep learning because it lets you:

1. Create and Visualize Models with Just a Few Lines of Code.

MATLAB lets you build deep learning models with minimal code. With MATLAB, you can
quickly import pretrained models and visualize and debug intermediate results as you adjust
training parameters.
 2. Perform Deep Learning Without Being an Expert.
You can use MATLAB to learn and gain expertise in the area of deep learning. Most of us
have never taken a course in deep learning. We have to learn on the job. MATLAB makes
learning about this field practical and accessible. In addition, MATLAB enables domain
experts to do deep learning – instead of handing the task over to data scientists who may not
know your industry or application.

3. Automate Ground Truth Labeling of Images and Video.

MATLAB enables users to interactively label objects within images and can automate
ground truth labeling within videos for training and testing deep learning models. This
interactive and automated approach can lead to better results in less time.

4. Integrate Deep Learning in a Single Workflow.

MATLAB can unify multiple domains in a single workflow. With MATLAB, you can do
your thinking and programming in one environment. It offers tools and functions for deep
learning, and also for a range of domains that feed into deep learning algorithms, such as
signal processing, computer vision, and data analytics.
With MATLAB, you can integrate results into your existing applications. MATLAB automates deploying
your deep learning models on enterprise systems, clusters, clouds, and embedded devices.
See more deep learning features and get example code.
Related products: MATLAB, Computer Vision System Toolbox™, Statistics and Machine Learning
Toolbox™, Neural Network Toolbox™, and Automated Driving System Toolbox™.
Deep Learning with MATLAB
MATLAB makes deep learning easy. With tools and functions for managing large data sets, MATLAB
also offers specialized toolboxes for working with machine learning, neural networks, computer vision,
and automated driving.
With just a few lines of code, MATLAB let you do deep learning without being an expert. Get started
quickly, create and visualize models, and deploy models to servers and embedded devices.
Teams are successful using MATLAB for deep learning because it lets you:

1. Create and Visualize Models with Just a Few Lines of Code.

MATLAB lets you build deep learning models with minimal code. With MATLAB, you can
quickly import pretrained models and visualize and debug intermediate results as you adjust
training parameters.

2. Perform Deep Learning Without Being an Expert.

You can use MATLAB to learn and gain expertise in the area of deep learning. Most of us
have never taken a course in deep learning. We have to learn on the job. MATLAB makes
learning about this field practical and accessible. In addition, MATLAB enables domain
experts to do deep learning – instead of handing the task over to data scientists who may not
know your industry or application.

3. Automate Ground Truth Labeling of Images and Video.

MATLAB enables users to interactively label objects within images and can automate
ground truth labeling within videos for training and testing deep learning models. This
interactive and automated approach can lead to better results in less time.

4. Integrate Deep Learning in a Single Workflow.

MATLAB can unify multiple domains in a single workflow. With MATLAB, you can do
your thinking and programming in one environment. It offers tools and functions for deep
 learning, and also for a range of domains that feed into deep learning algorithms, such as
signal processing, computer vision, and data analytics.
With MATLAB, you can integrate results into your existing applications. MATLAB automates deploying
your deep learning models on enterprise systems, clusters, clouds, and embedded devices.
See more deep learning features and get example code.
Related products: MATLAB, Computer Vision System Toolbox™, Statistics and Machine Learning
Toolbox™, Neural Network Toolbox™, and Automated Driving System Toolbox™.