0% found this document useful (0 votes)
21 views9 pages

Deep Learning Artificial Intelligence

Uploaded by

Syed Zafar Ali
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
21 views9 pages

Deep Learning Artificial Intelligence

Uploaded by

Syed Zafar Ali
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
You are on page 1/ 9

 Artificial Intelligence

 Applications

 Technology

 Deep Learning

 Editors Pick

 Guest Post

 Machine Learning

 Resources

Does deep learning always have


to reinvent the wheel?
By
Andreas Maier
-
January 4, 2019
0
1051
Photo Credit: Pixabay
Machine learning and in particular deep learning revolutionize the world as we
know it today. We have seen tremendous advances in speech and image
recognition, followed by application of deep learning to many other domains. In
many of those domains, deep learning is now state of the art or is even going
beyond it. A clear trend is that networks are growing more and more complex
and more and more computationally demanding.

Today, we are building ever increasing networks that are built on top of
previous generations of network topologies. As neural networks are inherently
compatible with other neural networks, we can combine and adapt them to new
purposes. If you aim to tackle a new problem, there are no clear guidelines that
define an appropriate network topology. The most common approaches are to
have a look at the work of others that attempted to solve similar problems or to
design an entirely new topology on your own. This new design is often inspired
by classical methods, but it is up to the network and the training data to learn
the correct weights such that they converge to a plausible solution. As such
they are even networks that learn well-known functions such as the Fourier
transform from scratch. With the discrete Fourier transform being a matrix
multiplication, it is often modeled as a fully connected layer. With this approach
it is immediately clear that the two disadvantages cannot be avoided: First, the
fully connected layer introduces a lot of free parameters that may model
entirely different functions. Second, the computational efficiency of a fast
Fourier transform can never be reached with this approach.

-Advertisement-

If we already know that a specific function is required to solve a particular


problem, it comes to our mind to ask the question whether it would not be of
advantage to include it into the structure of our network as a kind of prior
knowledge. The method of “precision learning“ investigates exactly this
procedure in a new theoretical framework. While this idea seems simple and
intuitive, the theoretical analysis also identifies clear advantages: First, the
introduction of a known operation into a neural network always results in a
lower or equal maximal training error bound. Second, the number of free
parameters in the model is reduced and therewith also the size of the required
training data is reduced. Another interesting observation is that any operation
that allows the computation of a gradient with respect to the inputs may be
embedded into a neural network. Even a sub-gradient is already sufficient as we
know from, e.g., max pooling operations.

Interestingly, this piece of the theory was only published in 2018. It was
developed for the theoretical analysis of embedding of prior physical knowledge
into neural networks. The observations also very nicely explain why we see the
tremendous success of convolutional neural networks and pooling layers. In
analogy to biology, we could argue that convolution and pooling operations are
prior knowledge of perception. Recent work goes even further: there exist
approaches that even include complicated filter functions such as Vesselness
filter or the guided filter into a neural network.

The theoretical analysis also shows that modeling errors in earlier layers are
amplified by subsequent layers. This observation is also in line with the
importance of feature extraction in classical machine learning and pattern
analysis. Combination of feature extraction and classification as it is done in
deep learning, allows us to synchronize both processes and therewith reduces
the expected error after training.

As precision learning allows the combination of classical theoretical approaches


and deep learning, we are now able to drive these ideas even one step further:
A recent publicationproposes to derive an entire neural network topography for
a specific problem from the underlying physical equations. The beauty of this
approach is that many of the operators and building blocks of the topology are
well-known and can be implemented efficiently. They were still operations that
are computationally very inefficient. However, we know from other solutions to
similar problems that particular matrix inverses or other less tractable
operations can be represented by other functions. In this example, an
expensive matrix inverse is replaced with a circulant matrix, i.e., a convolutional
layer which is the only learnable part of the proposed network. In their
experiments, they demonstrate that the proposed architecture indeed tackles
the problem that could previously only be approximately solved. Although they
only trained on simulated data, the application on real data is also successful.
Hence, the inclusion of prior knowledge also supports building network
architectures that generalize well towards specific problems.

We think that these new approaches are interesting towards the community of
deep learning that is going well beyond only modeling perceptual tasks today.
To us, it is exciting to see that traditional approaches are inherently compatible
with everything that is done today in deep learning. Hence, we believe that
there are many more new developments to come in the field of machine and
deep learning in the near future and it will be exciting to follow up on them.

If you think that these observations are interesting and exciting, we recommend
reading our gentle introduction into deep learning as a follow up on
this article or our free online video course.

Text and images of this article are licensed under Creative Commons
License 4.0 Attribution. So feel free to reuse and share any part
of this work.

Note: This is a guest post, and opinion in this article is of the guest writer. If
you have any issues with any of the articles posted at www.marktechpost.com
please contact at [email protected]

Share this:

 Click to share on Twitter (Opens in new window)

 114Click to share on Facebook (Opens in new window)114

 Click to share on Google+ (Opens in new window)

Related

Do We Still Need Traditional Pattern Recognition and Signal Processing in the Age of
Deep Learning?November 28, 2018In "Artificial Intelligence"

List of Deep Learning Books to ReadJuly 31, 2018In "Artificial Intelligence"

Top Deep Learning PapersApril 26, 2018In "Artificial Intelligence"

Shop Related Products


Deep Learning (Adaptive Computation and Mach…
$33.69$80.00

(182)

Practical Convolutional Neural Networks: Im…


$34.99

(1)
Neural Networks and Deep Learning: A Textbook
$45.45$69.99

(12)

Feature Extraction and Image Processing for …


$77.31$79.95

(7)
Deep Learning: A Practitioner's Approach
$23.19$49.99

(20)

The Deep Learning Revolution (The MIT Press)


$26.95$29.95

(10)
Deep Learning By Example: A hands-on guide to impl…
$38.21$39.99

(2)

GANs in Action: Deep learning with Generative Adversarial Networks


$49.99
Ads by Amazon

Previous articleModernizing the Supply Chain Management through the Blockchain Technology

You might also like