In the last tutorial, we gave you a brief history of neural networks before introducing to ‘deep learning’. Now that you are familiar with what it refers to, I suspect you’re wondering how useful it really is. What does one gain from learning about the complicated networks, the hardware, and all of this information? I will try and answer this question with the help of examples from the real world that demonstrate the uses of deep learning technology. I believe these examples will not only leave you with a sense of what deep learning technology is capable of but also leave you asking – what is it incapable of? So, let us get started on the uses of deep learning:
1. The first example that comes to my mind is the auto-captioning feature on YouTube. It is extremely useful for those with hearing disabilities and to those, who prefer to watch videos on YouTube with subtitles.
2. Deep learning has also been useful in enhancing airport security. Spotting malicious or illegal activity at a place like an airport can be challenging. However, due to the ability of certain machines to process and make sense of large volumes of “unsupervised” data, they are being used to tackle this difficult real-world problem.
3. Cybercrime is estimated to cost the global economy 400 billion dollars, the bulk of which is expected to be committed using fake credit card accounts. Artificial intelligence and deep learning have been identified as technologies that can prevent such a huge loss to the global economy, as is articulated in the following article.
4. A recommendation engine also referred to as a recommender system, is a tool that predicts what a user may or may not want out of a list of given items. Recommendation engines are becoming increasingly popular by the day, especially among e-commerce companies. According to a recent survey, 35% of the people who shop online do so because it is easier for them to find things online than it is in the traditional mode of shopping. The e-commerce industry stands to gain a lot from these recommendation engines, as they continue to widen their user base.
5. “There is nothing new in the world except the history that you do not know”. Some of the world’s history is buried in the graveyard of dead languages that have proved to be thus far undecipherable. The only way to reclaim this history is by figuring out what those languages mean. This was proving to be an impossible task for even some of the best of the human minds attempting the feat. However, with deep learning technology, machines have given a ray of hope to those engaged in reclaiming these long-lost languages and histories.
6. Another interesting use of deep learning is to bring black and white images back to life by infusing color into them. Trained by a million images from the Imagenet dataset, the algorithm is slowly but steadily learning to predict the colors that were greyed out in the original images.
7. 90% of the time we spend on mobile phones is used on emails or messaging. The companies that provide these services are beginning to include chatbots in their apps that interact with the customers via auditory or textual means. These bots are becoming increasingly popular among users as they are able to process unsupervised data in order to personalize user experience. An example of one such chatbot can be found below.
8. Charlie Chaplin was a creative genius, whose career spanned over five decades. During this time, he worked in several silent films that were as entertaining as his other films that included sound. If anything, these films were more fascinating due to their lack of audio that forces the viewer to imagine what the scene might sound like. What if I told you that you didn’t have to imagine this anymore? Deep learning has now made it possible for computers to add sound to silent films that were previously reliant solely on their visuality to communicate stories.
9. Probably the most popular of all deep learning tools is the one that enables people to translate words and sentences from one language to another. Speaking from personal experience, the tool was extremely useful for me in college when I was forced to cope with a higher level of English without much supervision. As a student, who was educated in a Hindi medium school, tools like google translate proved to be extremely useful in facing this challenging real-life situation. The tool, I am sure is also extremely popular amongst tourists, who find it difficult to communicate with locals in a foreign land. Deep learning is not only reducing the barriers of learning by making foreign languages more accessible but also bringing people closer by helping them communicate with each other despite linguistic differences.
10. A startup backed by Google called DeepMind Technologies is driven by the goal to use artificial intelligence in order to make the world a better place. Their projects focus on providing better healthcare to individuals around the world and tackling global climate change by reducing energy consumption. As a gamer, however, the project that caught my eye was the one that developed an artificial intelligence agent with the ability to automatically master 49 classic games created by Atari.
11. It’s fascinating to think of driverless cars. And when this thought becomes reality, it gets even more interesting. Google and Uber are competing with each other to turn this fascinating idea into reality. They’re doing their best to develop safe to use driverless transportation. Whether this technology is good or bad is a different discussion altogether – one that we shall leave for another day. For now, you may read a bit more about the company developing this new technology for Google.
12. Alzheimers, Parkinson’s and Cancer are some of the diseases that scientists have struggled to find a cure for. They have invested a lot of time and energy researching these diseases. And now, deep learning is helping them in their efforts. Check out the following link to know how deep learning is helping scientists accelerate drug discovery.
13. So far, I have given you examples of how useful deep learning can be. However, like everything else, deep learning too has its cons. The technology is still prone to misuse. For instance, it was used recently to fabricate a video of Barack Obama. The infamous video became widely popular on social media, despite being fake. The video was made using segments from existing video and audio clips of the former US President. Although deep learning technology can be sophisticated at times, there is no denying the fact that it can prove to be detrimental when misused, as is proved by this example.
14. It’s a proven fact that our visual memories are better than our verbal ones. Therefore, it might be easier for us to draw something we have seen on paper than to describe it in words. But what would be the use of our visual memories in the digital age, if we couldn’t use it to find things online. Luckily, researchers at the Queen Mary University of London recognized this problem and have created a solution. They have used deep learning to develop a technology that allows you to find objects using drawings.
15. Have you ever felt the desire to communicate more about yourself in any given instance? What if you could create a digital billboard around your face that revealed more about yourself? That’s what a startup called Blippar, which is committed to using technology in order to augment reality, allows you to do. Using their app, you can save a picture of your face on their servers, along with personal tidbits that you want to share with others. The others may access this information in the form of a halo around your head by scanning your face using the app.
16. Microsoft was one of the first companies to take on the challenge of speech translation. Several other companies joined this challenge later. Recent developments in deep learning technology have improved the quality of translations greatly. Skype is a good example of how advanced this technology has now become. Their speech translation feature provides real-time speech translations on their software.
17. In recent times, deep learning has been extremely useful in preventing loss of life and property due to its ability to predict disasters efficiently. Using a large amount of data that is available from satellites, sensors, expert knowledge, and hazard models, the technology has helped greatly in protecting people from natural calamities.
18. Traffic – who haven’t encountered this problem. There is nothing worse than being stuck in traffic, especially when you’re getting late for an interview, or for an important meeting. Luckily, a team of researchers, have now used deep learning to come up with, what they believe, is a permanent solution. Could this be it?
19. Being sick can be difficult. It is one of those times when you crave for constant attention and care. But in today’s day and age, attention can be difficult to get. Everyone seems to be busy attending to “important” matters. In such situations, artificial intelligence offers a new solution. Deep learning has helped the medical industry greatly. And it is being used now to revolutionize the provision of health care with the creation of personal assistants that promise to deliver the attention and care you need when you’re sick.
20. Lastly, much like human curiosity, deep learning knows no bounds. It is as useful in processing information about the earth, as it is about space. In recent times, deep learning has helped astronauts and researchers greatly in space exploration. They have been useful not only in capturing images and recording data but also in processing it and providing insights that have aided space teams in their decision making.
That brings us to the end of the article. I hope you enjoyed reading about how deep learning technology is being used to solve real-world problems. You may already have been aware of some of these uses, but I hope they gave you a sense of what is possible with this technology. I may have missed out on some interesting projects that are using this technology creatively. If you know of any such project, I urge you to share them in the comments section below. Make sure you subscribe to our blog to receive updates for the remaining articles in this series. Until next time, keep learning.
We hope this helped you know the importance of Deep learning technology in today’s world and to learn more, enroll for the Deep Learning course conducted by Acadgild.