Overview

In this post, we will be discussing the definition of Deep Learning, the general ‘neural’ way in which it operates. We will be seeing how it is quite relevant today and the growing interest in its theory, capabilities and possible future utilities. Then, we will be previewing a number of applications that Deep Learning is currently being in or are developed. We will be discussing the great potentials and certain limitations that this topic entails, and how is the future of deep learning might be looking like. Finally, closing with some final thoughts on the Deep Learning and its possible outcomes.

 

Definition

Deep Learning is, relatively, a new area in Machine Learning. It has been introduced and developed with the objective of moving Machine Learning closer to one of its original goals: Artificial Intelligence.

 

It vaguely models itself in a fashion that is similar to the complex and non-linear way the human brain categorizes and analyzes information and makes future predictions based on the learned data. Unlike traditional Machine Learning, Deep Learning attempts to simulate the way the human brain works, learns and processes information by creating artificial structures, called “neural networks”, that can extract complicated concepts and relationships from data, then build and extrapolate on it. Deep learning models improve through complex pattern recognition in pictures, text, sounds, and other data to produce more accurate insights and predictions.

 

Some good examples of Deep Learning frameworks are Apache MXNet, Microsoft Cognitive Toolkit, Torch. Other, newly, emerging ones are eightfold.ai, and missinglink.ai.

 

Why is it Relevant Today?

The fastly-growing importance of AI and Deep Learning can be seen, for example, by observing the investment value in the AI market. Venture Capital investment in AI doubled in 2017, attracting $12B compared to $6B in 2016, and is expected to reach $58B by 2021. Companies are increasingly turning to Deep Learning because it allows computers to learn independently and undertake tasks with little, or no, supervision, promising extraordinary abilities benefits for science, industry, communication, and security.

 

For a good reading on its relevance and importance, check this article.

 

Use Cases Examples

Deep Learning applications are all around the place, I’ll preview a few of them here.

 

Automated Driving: Companies building self-driving cars, like Tesla and Google, are basically teaching a computer how to take over key parts, or all, of the driving process, relying on its experience in interpreting and reacting to the digital sensors system attached to it. Researchers are using deep learning to automatically detect and appropriately interpret objects such as stop signs, traffic lights, and pedestrians.

 

Health Care: From breast or skin cancer early detection methods to creating personalized medicine on the basis of a Biobank’s data, deep learning will be reshaping the way our health systems function. Innovations in AI are advancing the future of precision medicine and population health management in extraordinary ways. Computer-aided detection, quantitative imaging, decision-making support tools, and computer-aided diagnosis are some of the big focuses of in the field.

 

Aerospace and Agricultural: Finding the optimal routes for spacecrafts and satellites and the precise maneuvers to take, optical analysis of forested or planted lands for early detection of diseases and fires, locating icebergs by analysing satellite images of the path of ships, to identify, locate and analyze objects of interest over massive databases of satellite images, and many more are some of the applications currently being researched on.

 

Electronics: Deep learning is being already widely used in automated hearing, speech and text recognition, translation, and computer vision. Alexa, Siri, Google Now and Cortana are few examples of some of those daily encounters people have with deep learning applications.

 

Future?

The prospects are intriguing, clearly, as companies like Google, Microsoft and Facebook are spending millions and millions on research into advanced neural networks and deep machine learning. This comes with numerous promising projects and applications that would reshape the way humans interact with AI at daily basis.

 

In the meantime, significant hardware and algorithmic developments have been underway, building up to what appears to be a race of new applications for deep learning frameworks in areas as diverse as energy, medicine, physics, and beyond.

Some of the bright current research projects are working on spotting invasive brain cancer cells during surgery reducing the error margin critically, Restoring the colors of Black and White pictures and films, Pixel restoration for faces of low resolution producing semi-original versions and many more.

 

Closing Thoughts

We have seen what Deep Learning is, and what some of the valuable technologies that are driven by some of those ‘neural networks’ can do and what, certainly, great powers some of them have. But that doesn’t mean that we understand much about the exact way in which those deep multi-layered systems work or can be made to necessarily converge to a solution within a finite time.

Humans understand the tremendous potentials in embracing AI that could possibly become smart and powerful enough to become part of our life. However, people must be aware of what’s wanted from these smart machines, as they are not solely serving our interests, at least we have no guarantee of that so far.

 

In conclusion, we have gone through the definition of Deep Learning, the way it works and its growing importance and popularity among investors, tech companies and new startups. Later on, we went into a thread of various applications and some possible use cases. Concluding with some final thoughts on certain limitations in terms of ambiguity and complexity.

References

Deep Learning

http://deeplearning.net/

 

What is Deep Learning? 3 things you need to know.

https://www.mathworks.com/discovery/deep-learning.html

What is deep learning? Everything you need to know

https://www.zdnet.com/article/what-is-deep-learning-everything-you-need-to-know/

 

25 Machine Learning Startups to Watch 2018

https://www.forbes.com/sites/louiscolumbus/2018/08/26/25-machine-learning-startups-to-watch-in-2018/#7c202c266f99

 

30 Amazing Applications of Deep Learning

http://www.yaronhadad.com/deep-learning-most-amazing-applications/