Welcome to NeeshAi!
In the last tutorial we introduced you to a brief history of neural network and also gave some idea about “Deep Learning”. So now that we are familiar with the concept, questions sparking your mind should be- what is the use of all this? All this complicated networks, lots of hardware, plethora of data, do we actually gain anything substantial and informative from it? I will try to answer all these questions while we proceed through the write-up but first let me promise you one thing –If at the end of the tutorial your question does not turn 180 degree and translate into- “Is there anything of some sort that Deep Learning cannot do?” the very motive of mine to write this tutorial will be defeated.
To start with, these are couple of things that I find quite amazing or to say magical. Please note that I may not mention some of the applications that you might find pretty cool or amazing because of the fact that maybe I do not find them quite fascinating or maybe I am just not aware of them. Ok, so let me start speaking without wasting your time any further.
- The first application that comes to my mind is “Auto-captioning of YouTube”. This feature is literally life changing for those with hearing disabilities; aside from those who prefer to read lectures/videos simultaneously while watching it.
2. Airport security is another big thing that I can think of and there is no point emphasizing the fact how important is it?
3. Cybercrime is estimated to cost the global economy 400 billion dollars and credit card fraud accounts for a large proportion of this cost. Deep learning seems quite promising in this area. It can detect the frauds and save global economy from such a huge loss.
4. A recommendation engine also referred to as a recommender system is a tool that predicts what a user may or may not want among a list of given items. Recommendation engines are becoming popular day by day among e-commerce companies. According to a survey 35% of the people who buy online do so because they find that it is quite easy process compared to traditional method of shopping. As more and more people will have internet access, e-commerce industry is going to gain a lot from these recommendation engines.
5. “There is nothing new in the world except the history that you do not know” and there are various things in the graveyard of history about which we have no idea. The only believable sources to know the history is decipher those languages and read them. The human mind was not capable of doing so and thus history was history forever. But with the emergence of deep learning it seems very much possible now.
6. Our brain is capturing each and every moment of our lives and extracting meaningful information from it. It took a lot of experience for us to finally identify subconsciously that sky is blue and grass is green. Let’s suppose you ask neural networks to do the same thing –what do you think how much time it will take to learn it? It seems quite difficult is not it? But deep learning has made it possible now.
7. Around ~90% of our time on mobile is spent on email and messaging platforms. Every company would like to make something for the platform where their potential customer spends most of his/her time.
Chatbot provides companies such an opportunity. Chatbot are messaging apps that we interact with via a chat interface and which gives us real human to human chat apprehension.
8. I will be amused if someone reading this blog has not heard the name of Charlie Chaplin. Consider for a moment how amazing it will be if there was some mechanism through which we could add sound effect to all those movies. Oh wait, it is possible already and it’s truly amazing.
9. When I came to college it was hard for me to cop up with all the English words that were thrown at me because all my life I was a Hindi medium student. I wished I could find something that could translate all those fancy words into Hindi and guess what- there was something.
10. Google-backed startup DeepMind Technologies has built an artificial intelligence agent that can learn to successfully play 49 classic Atari games by itself. How amazing is this!!
11. Is not quite fascinating to even imagine a car running on the road without a driver and it becomes more fascinating when it turns into reality. Google and Uber are competing with each other in the field and developing safe and easy transportation vehicles. Whether it’s a good thing or bad thing is a whole different issue and we will discuss it probably some other day.
12. Alzheimer’s. Cancer. Parkinson’s just to name a few diseases scientists are struggling to find the cure of. It takes year and year to discover a particular drug. Now with the help of deep learning it is quite possible to discover a drug within few months. The cure for Alzheimer disease is believed to be discovered with the help of deep learning.
13. Till now we have seen quite a few cases where deep learning has proven to be quite useful but everything has two sides and deep learning is no exception. Just to give you an idea-few months back a video of Barak Obama was circulated and people believed it to be true. It turns out later that video was fake and was generated using deep learning on existing video/audio clips of Obama. While it may seem quite fun for now, we cannot deny the fact that it has the capability of destroying somebody’s careers.
14. It’s a proven fact that our visual memory is way better than reading memory. Sometimes it’s hard to describe something in words than just drawing it on the paper. Suppose you have seen something and you don’t know what it was but at the same time you want to buy it or you want just to find out the name of that particular object. Researchers at Queen Mary University of London created a tool that recognizes hand-drawn sketches on a smartphone or tablet and uses them to search for products online. How cool is it?
15. Just imagine-what if you could turn your face into a digital billboard? The augmented reality startup Blippar allows you to scan your face with its app and it saves a version of you on its servers. You can then fill out some personal tidbits. In future if someone scans your face with the app, he/she can see that information as a halo of bubbles floating around your head.
16. Microsoft was one of the first to delve into the challenge of speech translation. Recent improvements in deep learning have made possible better and better translation outcomes. Skype is the latest example of this; it’s a speech to speech translation which works on the real time basis.
17. Probably the most useful use of deep learning till date has been disaster prediction. An abundance of data is available from satellites, sensors, expert knowledge or mathematical hazard models. Deep Learning is using these data and making sense of all this in a timely manner and accurately predicting the upcoming disasters and thus helping in mitigating or reducing the loss of life and property.
18. We all have faced the traffic one or another time in our life. It’s a problem we can all relate to. Nothing can be worse than waiting for a green light while you are late for office or somebody is struggling between life and death or any other urgent thing for that matter. Team of researchers has presented an 8 lane model that can make go away the problem of traffic forever. We all are waiting to see it as a reality soon.
19. Imagine your father lying on a bed in a hospital. Doctor has said that he has to stay some 15 days in hospital itself. Though you care about you father more than anything else, due to some urgent requirement you cannot be with him all the time. Deep learning is helping medical industries in a great way from detection to diagnosis to personal assistant.
20. The space far-far from the earth also is not beyond the scope of deep learning and artificial intelligence. It helps astronauts and robots make swift decision and plan quick at the same time helping in exploration, image capturing and processing.
I am hoping you had a great time reading this article. Some of you knew these applications some did not but at the end what matters is we all have touched the surface of what a Deep Learning model is capable of. If you think we have missed some of the applications that should have been there in the article, do let us know in the comments section below. We are going to release the next parts in the series “Implementation of a Neural Network” very soon. Stay tuned with us.
Do share your feedback and follow NeeshAi to learn more about theoretical and practical blogs on Tech! Keep learning.
References:
