Revolutionizing AI with Deep Learning – A game-changing Technology

World Economic forum says about Deep Learning powered AI as “We stand on the brink of a technological revolution that will alter the way we live, work and relate to one another. In its scale, scope, and complexity, the transformation will be unlike anything humankind has experienced before”.

What is Artificial Intelligence?

Artificial Intelligence is a technology to simulate human intelligence in machines by making them think, solve problems and respond by learning like humans. 

Once, Artificial Intelligence was science fiction because of the way the machines were trained with step-by-step instructions (billion lines of software code). Now it is a reality    –Thanks to changing in approach by teaching machines, called “Deep Learning”–A Technique for Implementing Machine Learning.

Machine Learning, importantly Deep Learning technology, is driving Artificial Intelligence to the next level.

What is Deep Learning?

Before diving into Deep Learning, let us quickly understand Machine Learning (ML). ML is the practice of using algorithms to parse data, learn from it, and predict. So instead of programming software for a specific problem, train the machine using enormous amounts of data and algorithms that give intelligence to machines to learn how to solve the problem.

Machine Learning algorithms almost always require structured data, while Deep Learning networks rely on layers of Artificial Neural Networks (ANN).

Wiki defines as “Deep Learning is a class of Machine Learning algorithms that use multiple layers to progressively extract higher-level features from the raw input.”

In English, it is teaching machines in the same way that we teach humans. Let me brief you with an example. 

While driving, you can easily differentiate between a car and a truck. You can easily differentiate between a cargo truck and a fire truck. Our brain is tuned to give way to fire-truck without even noticing we are doing it. This is because over the period our brain is trained to respond so. But for machines, “a truck is a truck is a truck”. It was difficult to program the machine to distinguish and respond to the way humans respond. 

With Deep Learning techniques, technologists teach computers the same way they teach humans. Let me help you to get the basic idea behind Deep Learning. You are driving your daughter who just learned the word “truck”. Your daughter will point at every vehicle passing by and say “Truck!”. You cheer her and perhaps give her a little more information. You may say, “Yes honey, it is a Truck, actually “fire truck”, and start explaining what is a fire truck and how we should respond and why we respond so”. In case if she points a car and says the truck, you tell her that she is near, but that is something different called “Car”. Over the period your daughter learns the difference between a car and a fire truck and a cargo truck. She also learns how to respond to each type of vehicle.

Similar techniques are used to teach machines, too. There may be a billion of such automobile objects. The trainer (programmer) starts by giving a machine the basics of automobile objects and Images of all kinds of automobiles with its basic info–More images, better learning. This gives the ability to the machines to distinguish between a car and a truck as humans do.

Using this technology as the backbone, On October 20, 2016, Otto and Budweiser completed the world’s first shipment by self-driving (No Driver), 18-wheeler truck that drove with a crushing speed of 55 mph on I-25 from Fort Collins, through Denver to Colorado Springs for about 120 miles. Meaning, the Deep Learning engineers trained the truck to see and respond like humans.

Here is another example of how Deep Learning technology helps to treat diabetes that causes blindness. There is a drug that can prevent it. However, it only works, if it detects the warning signs very early. Now the machines that were deep learned can look at the pictures of diabetics’ patient eyes and alert early blindness warning signs.

Few more applications of AI-powered by Deep Learning:

• Helps the farmers to increase the crop yields by separating weeds and good plants so they can spray only the weeds with herbicide.
• Facebook applying Deep Learning to automatically recognize people in the picture. 
• The Chinese Communist Party (CCP) uses this technology for mass surveillance and
  
  build Social Credit System (SCP) by profiling their citizens (Ref: China Surveillance Wiki) 
• Boeing uses Deep Learning in the airplane maintenance to detect mechanical failures before a part fails
• Over the period Alexa’s self-learning capability can read the news that an individual interested-in
• IBM Watson can read medical images and help in patient diagnosis
• Helps Finance Industry to identify suspicious transactions
• Deep Learning can provide potential treatments for Ebola & Covid-19, multiple sclerosis and various types of cancer

The applications of Deep Learning are practically limitless. Here is the bottom line:

Few years before a company isn’t competitive if it is not on the internet. Today all companies are on the internet, whether they are a small company working from a garage or a trillion-dollar company that has gigantic lavish offices in every country. Only very few companies started leveraging the power of Artificial Intelligence and Deep Learning technologies now. Within a few years from now, every company must embrace this technology or they will get diluted. 

Potentially there is an enormous market opportunity for Artificial Intelligence and Deep Learning. The leading research companies (McKinsey), top market advisers (Gardner), technology influencers (Jeff, Bill, Sundar & Elon) are predicting as 20 trillion-dollar opportunity.

Artificial Intelligence and Deep Learning need an enormous number of data and the companies that have most data can provide the fuel that feeds Deep Learning and those companies will be a front-runner in this game. Now the companies know how to train the machine right and have access to a vast amount of data to train the machine, It’s all about speed now. Because the faster you train and feed the machine, the faster it can make advancements. So, the industry needs groundbreaking processors that can process huge data, learn and respond 1000 times faster than they could previously do. Long back, Deep Learning has primarily seen as software programming. Start from the year 2016, the need for more efficient hardware (AI accelerator) necessitated by the industry to speed up AI/ML/DL implementation.

As defined by Wiki, “An AI accelerator is a class of specialized hardware, integrated circuit chip that contains billions of MOSFET transistors with in-memory computing capability, accelerator or computer systems designed to speed up Artificial Intelligence applications, especially artificial neural networks, machine vision, and Machine Learning. Typical applications include algorithms for robotics, internet of things and other data-intensive or sensor-driven tasks” (Ref: AI Accelerator) )

Conclusion

Any startup investing in AI accelerator will be the next Microsoft, Apple, Google, or Amazon. Being part of this next technology revolution is not optional. You can be a Technologist or an Entrepreneur or an Investor. Shying away only puts you as a long-gone individual.

Forbes says that AI and Deep Learning will virtually change the economics of every industry. 

Let’s see, who will lead this game?

Cheers,