Artificial Intelligence is not a single technology; rather, it’s an umbrella term for technologies inspired by human biological systems. AI constitutes a variety of software, algorithm-driven approaches with data, that together simulate any number of human cognitive functions.
To the lay person this is “dare to dream” stuff – but the truth is that AI has been around since the 1950s. In the early years, there was a somewhat naïve view regarding what it would take to build an “intelligent” machine, so progress was slow. Consider the tools computer scientists had to work with at that time: Programming required very detailed instructions and rules to get the computer to do anything useful. Computers filled entire rooms, and the Internet was just a thought in the back of someone’s mind.
Fast forward to today: Over 1 billion people and millions of computers and sensors are interconnected across the world. Computers have shrunk – and many are not much larger than a grain of sand. A convergence of forces over the last 5-10 years has enabled a virtual explosion of possibilities for AI that is quickly moving into every facet of how we work, live, and play. The most influential drivers include:
• Availability of data
• Cheaper and faster hardware
• Better algorithms
Quality training data (text, image, audio, user activity, or knowledge graph) is the new competitive advantage in AI.
As its name implies, artificial intelligence is about replicating what the human brain knows. The learning is iterative, and requires processing thousands or millions of examples of something (like photos, music, words in a dictionary, etc.) to become competent in a task. The level of data available significantly impacts the speed of learning, and the ultimate competence AI can attain. For example, Google’s success rate in delivering more precise results in website and image searches is reliant on the unprecedented volumes of data fed into Google’s AI algorithms.
The Big Data generated as a byproduct of our increasingly interconnected existence is a vital resource in AI development. Businesses are not only able to collect, analyze, and filter specific information at an amazing speed – but also recognize patterns that drive decision making policies based on this analysis. Thus, AI provides the missing link – intelligence - that helps businesses further their goals.
New types of algorithms lie at the heart of the new AI wave and it’s safe to say that without these, the data and computing power would amount to nothing. It might come as a surprise that it is all based on 1950s-era technology known as the Artificial Neural Network (ANN) - an attempt to create a computer model of the network of nerve cells in a human brain. Loosely speaking, a neural network is an interconnected web of artificial neurons that either fire or not based on the input to the neuron.
A key part of building these neural networks is to train them to do the correct thing when they see data. The study of these algorithms has now spawned its own sub-field of AI known as deep learning, a reference to the number of neuron layers in the neural networks. We’ll talk more about deep learning in the next section.
The article has been penned down by Allan Frank, Chief Scientist, Capgemini