Artificial intelligence is one the most talked about and promising technologies of today. It seems like we hear about the new accomplishments in the AI field and stunning results Sophia manages to achieve on a daily basis. But just as we’ve seemed to have grasped the concept and come to grips with both the amazing results and constant warnings against AI from Elon Musk, it becomes more developed and two new concepts come to the scene – machine learning and deep learning. Though often the three terms are used interchangeably they don’t mean the same thing.  

Today we set the record straight as we explain each of the difference — deep learning vs machine and what we need to know about them. But let’s start from where it all began — artificial intelligence.  

Artificial Intelligence in a Nutshell

As you can see from the picture above, Artificial Intelligence is a more general category that encompasses both Machine Learning and Deep Learning. Basically, AI implies any machine that is capable of demonstrating any intelligence that leads to solving a certain problem.

The term Artificial Intelligence goes way back. It was first introduced in 1956 during Dartmouth Conference organized by American computer scientist John McCarthy who is now considered the “father of AI”. From then the interest in AI has only been on the rise and new centers to explore the potential of the technology. Its first notable success was in 1951 when Ferranti Mark 1 machine used an algorithm to master checkers. In the 70s the world saw the first intelligent anthropomorphic robot, WABOT-1 which was built in Japan in 1972. And in 1997 IBM’s Deep Blue computer beat then a world chess champion, Garry Kasparov thus proving that machines can not only lead intelligent conversations or perform human-like actions but also actually be superior to human beings in certain aspects.

Generally, when speaking of artificial intelligence, it can be viewed in two aspects: general and narrow. General AI deals with machines that much like humans are capable of tackling diverse tasks. The advanced general AI machine is the one that has cognitive abilities and a general understanding of the world, we humans live in, topped with the ability to process this data much more quickly and efficiently than us humans. While narrow AI has a more limited scope of abilities. They can perform a very specific task but usually a lot better than a human.

What is Machine Learning

Deep Learning vs Machine Learning

A subcategory of artificial intelligence, machine learning is based on the idea that machines can learn from data, identify certain patterns and act upon this information. The whole concept is based on the idea that machines can “learn” from the data they’ve got and then make an accurate prediction based on what they’ve learned.

One of the most popular examples of machine learning from our everyday life would be Netflix recommendation system. 80% of what we watch on Netflix is recommended to us. To accurately predict what their users might want to watch or don’t even realize they want to watch is thanks to machine learning and an algorithm Netflix is using. The data the service feeds its algorithm can be divided into two types: explicit and implicit. Explicit data is what users tell the streaming service themselves, for instance by liking a certain TV show. Implicit data is based on user behavior. If for example you binge-watched The Crown in a weekend, you didn’t exactly tell Netflix you loved it but the algorithms understand it behaviorally. Most of the data Netflix processes is implicit. It later uses it to split their users into over two thousand taste groups. Depending on the group you’re in, you will get your recommendations.

But that’s not all machine learning is capable of. It can act as a virtual assistant, and be quite good at their job, predict traffic, or personalize your social networks feeds. But perhaps one of the most interesting and useful applications of machine learning is Computer Vision, a technique that teaches machines to extract useful information from images and videos. It’s the technique Pinterest uses to scan images for objects and thus suggest similar pins for users. Computer Vision is currently applied for various purposes across different industries like healthcare, architecture, automotive industries, insurance etc.

Deep Learning vs Machine Learning

Deep Learning vs Machine Learning

Now on to Deep Learning, a further evolution of Machine Learning. When comparing Deep Learning vs Machine Learning, the two terms are often mixed up and used interchangeably, which is not that surprising given that they both act similarly, but yet with a significant difference.

Much like Machine Learning, Deep Learning is designed to analyze data but it has a notable distinction — it does so using artificial neural networks that, in a way, resemble the neural networks of a human brain. Machine Learning typically focuses on pre-programmed features that are brittle and not scalable, while Deep Learning tries to learn features directly from data as opposed to relying on those hand-engineered by a human. Deep learning has exploded in the recent years and, thanks to the recent advancement in computing, can be applied very complex problems.  Deep Learning at this point is the most human-like version of Artificial Intelligence

As to the impact of Deep Learning on our lives, it is already immense and will only grow in the future. Deep Learning is the key technology behind self-driving cars. It is applied for translation and language recognition. Deep Learning can even discern dialects of a language and learn it, also without the involvement of humans. It can add color to black and white photos and videos, as well as sound. In this instance, the machine associates the video images with a database of pre-recorded sounds and adds the most suitable one to go with what is happening in the scene. And this is not nearly the end of the impressive results of Deep Learning.  

The Future of AI: Where is it All Going?

Deep Learning vs Machine Learning

And finally, the most important question — where is Artificial Intelligence going? Should we be worried about what Elon says and will machines take over the world?

The experts consider Artificial Intelligence a great tool to amplify the effectiveness of humans, their productivity but on the other hand, it can pose a threat to humans and their capabilities.

The optimists believe Artificial Intelligence to be beneficial to healthcare as it can be applied to diagnosing and treating patients, helping senior citizen to lead a better quality of life or contribute to public health research.

Already, it is revolutionizing the workplace by increasing productivity, accuracy and efficiency. “No-collar” workforce is a massive trend in the digital world today. The idea is that AI-powered machines are designed not to substitute human resources but make their lives and work easier and more productive by freeing them from easy and repetitive tasks. This will allow humans to focus on tasks that require emotional intelligence. AI co-workers have already been successfully adopted by NASA who introduced a bot, George Washington. George works in HR and helps them by processing job candidates. George has several fellow bot-team members that help various departments.

The faith and the potential of this field are so high that the U.S. government has recently signed an order to prioritize Artificial Intelligence in its research and development spending in 2019.

In this blog post, we tried to clear the confusion surrounding Artificial Intelligence and compare Deep Learning vs Machine Learning in an effort to explain how they work, differ and are applied today.