Artificial Intelligence has over the years witnessed unprecedented advances which have proved to be continuous processes. AI has currently moved from research concept to engineering application. From self-driving cars to advanced military hardware and to the devices we interact with during the course of our day, artificial intelligence is gaining grounds. AI is concerned with having computers gather data and understand the world the way we (humans) do. The field is gaining a lot of momentum recently. Almost all leading I.T companies are now devoting significant amount of resources to AI research and development. Intel (one of the world’s leading chip manufacturer) recently bought an AI company (MobileEye) which specializes in design of chips and cameras for cars and trucks. At $15.3 billion, this is Intel’s second largest acquisition. Microsoft and Google are investing heavily in AI also.
AI has so many applications: in entertainment; in Engineering; in business; in medicine and in education (a few but to mention). Computers are nowadays used to predict election results, business success and stock index value. AI is at the center of all this. Many businesses now invest heavily to create Business Intelligence systems in order to get intelligent insights which give these companies an edge over their competitors. There have been several developments in the field which started almost 50 years ago. It is almost impossible to discuss all of these developments, however, the most recent ones will be discussed under the following headings.
The goal of AI is to create intelligent agents –machines that can think and take actions. Therefore, machine learning at the heart of AI. With recent advances like reinforcement learning, deep learning and response optimization, AI is making computer science exceed our expectations. Through machine learning, the system basically learns by itself using a lot of computations and data. In 2015, Microsoft and Google beat the best human at image recognition for the first time. Reinforcement learning is concerned with making computers learn and make decisions – through experience- with or without any prior knowledge of how its actions will affect its immediate environment. Deep learning is also on the ascendancy. Through Deep learning, computers work out how to extract general rules from huge amounts of noisy data by repeatedly applying complicated mathematical techniques. Only recently, a machine (AlphaGo) was able to defeat a human (Professional) player in the game of ‘Go’. This was hitherto impossible. The actual story of how it happened is long and interesting. However, to sum it up, AlphaGo used deep learning to beat the game’s best player. Go is (somehow) a complicated game. Even the game’s best players find it difficult to explain how they win. And this makes it extremely difficult to program a computer to play the game to the extent of becoming perfect. AlphaGo was programmed using the same AI techniques that older systems like Deepblue were built from, but the big idea is to combine them with new approaches that help the computer discover techniques for itself. This was achieved through Reinforcement learning.
The way neurons in the brain work inspired Deep learning which has already proved incredibly useful. The AI technique is being used in the processing of medical images and sieving through huge amounts of medical data. Last year a demonstration was made by a team from Google which showed that deep learning can be used to automatically diagnose eye diseases . A similar team from Mount Sinai Hospital (New York) used the same approach to analyze the electronic health records of patients’ and predict, with (an amazingly) high accuracy, what disease a person would go on to develop. Again, in 2015, Microsoft and the China University of Technology and Science taught a computer network how to take an IQ test. The system ended up scoring better than a college postgraduate.
In the areas of computer vision, tremendous success is been recorded too. With a somewhat low error rate, computers are now able to recognize good-quality printed text. Low quality (and perhaps old) documents maybe however difficult to recognize. Furthermore, current generation of ATMs can read handwritten checks . A program was implemented a while ago that identifies the category of a bird (duck, heron, hawk, owl, songbird) and recognizes major components of the bird’s body in pictures of birds. The program achieved a precision of about 50 percent on finding the components, and about 40 percent on identifying the category. Similar intelligent programs have also been implemented to identify images that represent simple phrases and those that extract events based on news articles. The former had precisions ranging from 5 percent to 85 and 100 percent for different images and phrases.
Furthermore, depth-sensing cameras are now able to capture objects in 3-D. Such cameras are implemented in industrial vision systems used in farming which can literally separate grain from chaff. Although there’s a huge gap between these advances and their ‘becoming perfect’ (as humans), it is however a significant success so far.
AI In Automobiles
Self-driving cars are becoming common and by the day more sophisticated. The number of AI systems in our automobiles is increasing rapidly. They do all sorts of things from improving driving performance to avoiding collisions on the road -by sensing nearby objects- and those sorts of things. New generation of cars nowadays come equipped with virtual copilot that is smart and intelligent . The system ensures safer driving by keeping track of the driving environment- the road, whether condition and surrounding objects. In the event the system perceives any danger, it automatically takes over from the human driver to apply evasive actions. Many elements of computer vision and machine learning are used here. Although the autonomous driver is never as perfect as the human driver, significant researches and trials are been conducted to perfect the technology. In the nearest future, self-driving cars will be common in many of our roads.
AI is having relevance in almost all aspects of human endeavor -business, learning, medicine, entertainment etcetera. There are various aspects of commercial use of AI. It is used for advertising, web searches and business analysis and forecast. It wasn’t until the last few years that AI could do things that people can’t do. Several milestones were achieved in 2015 in particular that made it possible for us to use it in all kinds of areas. With the world becoming more connected like never before through Internet of Things (IoT) and other similar technologies, it is believed that AI will be implemented in almost all the systems that we interact with –our toasters, cars and computers. All of these devices will have some form of (artificial) intelligence that will make them require little or no human input. In summary, AI has the potential to transform the way we live in unprecedented ways. Despite the advantages the field brings, the greatest fear (as argued by some scholars) is that continued research in AI will only result in creating sophisticated machines with (close to) perfect thinking ability as humans. Such machines could pose a threat to the entire human race. In a piece coauthored by Stephen Hawking it was argued that “Whereas the short-term impact of AI depends on who controls it, the long-term impact depends on whether it can be controlled at all” . However true or not this thought might be, AI is gaining grounds and seems to be awesome.