The machine learning technology proves that today’s computers are acquiring ever more astonishing skills. Perhaps their ability to take in and process newly acquired knowledge will one day enable them to sustain artificial intelligence. What exactly is the computer’s ability to learn?
We are naturally delighted to see babies smile when they recognize our faces as we lean over them. We are amazed to see little ones utter their first words as they repeat after us. Then comes the time when a child begins to make up stories about the fairy tale characters we have read to it about. Finally, it learns to count, draw, write and correct its first school paper. And then one day we hear it express opinions about the world. Opinions that are often so extraordinary they leave us in shock. At this moment, we are witnessing a human being becoming intellectually independent.
Machines that smarten up
It is not only people that learn from experience by acquiring knowledge from the outside world. We are now witnessing a time when sophisticated information technologies are also emerging from infancy. One development associated with Artificial Intelligence that captures the imagination of the entire world is machine learning. The name itself hints at a field of fully automated processes that rely on intelligent data processing and smart decision-making. AI is where the most ambitious R&D work is conducted in today’s world.
Machine learning, or limited self-sufficiency
Machine learning is a field positioned on the borderline between mathematics, statistics and programming, i.e. information technology. Its goal is to create complex algorithms capable of reaching optimal decisions and, even more importantly, continuous self-improvement. The algorithms that underpin machine learning are specific and highly sophisticated. By and large, they rely on a dynamic model that processes inputs (data) to make specific decisions. What is significant is that such algorithms have the ability to “self-learn” as they actively process the datasets that are entered. However, the entire mechanism has one serious limitation. And that is, as the computer executes its tasks, it draws on experience of the “supervisor”. What it means is that man – a programmer, operator or teacher – critically influences the way information is processed. His or her job is to support the machine by entering data batches, manually checking the conditions that result from analyses and remove system blockages. The computer’s self-sufficiency therefore continues to be limited as it depends on an expert. The general consensus is that the first people to witness machine learning were the IBM experts who tried to develop algorithm to help chess players improve their game. A landmark along the path came with the development of the Dendral IT system at Standford University in 1965 which automated chemical analyses. It is now recognized that the research led to the first computer-discovered compounds.
Some the latest research seeks to eliminate or at least strongly reduce the teaching role played by humans to ensure that algorithms learn independently.
Time for unlimited self-sufficiency
One of the most intensely explored areas today in the field of machine learning is deep learning, viewed as a subcategory of the former. Extensive mathematical structures that support multi-strand processing, referred to as neural algorithms, are capable of making decisions, correcting them by learning from their mistakes and, based on prescribed models, selecting from the available sets the data that most accurately addresses a given question or problem. In other words, they can learn independently using the deep learning method. Deep learning supports e.g. voice recognition, natural language, translation from various languages and image recognition. All these functions are particularly interesting to corporations such as Facebook and Google.
Deep learning systems are designed to emulate the workings of the human brain with its powerful neural networks. Such networks are what allows computers to learn independently without human supervision. In essence, such systems comprise large numbers of parallel processors fed massive amounts of data, including specifications for complex relationships between data. The networks use unlimited flexibility as neurons can be combined into layers. Every layer provides the next one with the results of the previous one until a task is solved.
By and large, a computer’s ability to learn, i.e. improve its operation, results from a single cause. Neural networks are collections of interconnected nodes. Each time a computer acquires a new experience or takes action, its connections reorganize themselves. They become ever more perfect allowing the machine to perform its assigned tasks more efficiently. The fact that computers adapt independently through the experience they gather rather than through a programmer’s interference was a breakthrough in the development of IT and its commercial applications. The main driver of interest in neural networks is unabated demand for structuring and searching through information. In the Big Data era, deep learning technology is our great ally.
Watson will tell you the whole truth
This may sound odd giving one an impression we are discussing equipment hidden away in military laboratories. But that is hardly the case. Deep learning is moving out to the streets and into our homes. The IBM-made computer, known more appropriately as artificial intelligence IBM Watson, can analyze huge sets of data. This incredibly efficient machine shows the possibilities brought to users with the advent of machine learning and deep learning mechanisms. Computers understand the questions asked in a natural language. As they search for answers to such questions, they go through a variety of vast datasets having to do with business, mathematics, medicine and law. IBM Watson relies on the ability to increase its capacities with each successive task. The more data the machine absorbs and the more tasks it receives, the greater its analytical and cognitive abilities.
The proliferation of such machines is precisely what today’s business and medicine are counting on. For that reason, any self-respecting technology corporations is making every effort to employ the solutions adopted for this ultra-efficient computer as broadly as possible. Similar solutions are being developed by Elon Musk as well as Google, Facebook and Microsoft.
How to win in the Chinese game of “Go”?
A curiosity as well as a good illustration of the above mechanisms and differences between machine learning and deep learning is the famous story of a computer defeating masters in the Chinese game of “Go”. Its rules allow for an almost infinite number of combinations of stones on the game board. (Compared to “Go”, chess is a simple game). First and foremost, the Google-developed algorithm Alpha Go enabled a computer to analyze millions of games played by humans (a case of supervised learning from a limited dataset). Much greater impact on the computer’s success came from the machine’s ability to analyze another set of situations unfolding on the game board – these resulted from the machine playing against itself (deep learning)! Needless to say, such work was only possible with the use of powerful servers and the cloud technology. It is nothing other than the availability of processing power that paved the way towards the incredible success of the Google algorithm.
We are all benefitting
The impact of intelligent machines is bound to be particularly profound in a range of areas. Tesla’s autonomous vehicles would be inconceivable without the technologies I am describing. One of the reasons why self-driving vehicles can navigate roads is their ability to read and analyze the images they encounter. Machine learning also has implications for the way information is structured. Consultancies that conduct their studies with the use of hundreds of millions of data points gain powerful analytical tools. Another likely beneficiary is the medical world Any instrument that allows medicine to model the behavior of the human body is worth its weight in gold. In another article, I described a contemporary client who demanded that the global market respond with a fully individualized services that are available on a moment’s notice. Learning machines are just the means to make that possible. Large Internet-based retailers such as Amazon use state-of-the-art algorithms to present their offerings on clients’ monitors in ways that are both better and faster.
Humanity’s collective knowledge at one’s fingertips
As mentioned earlier, the commercial implications of the technologies in question are increasingly more visible for the average user. Thinking machines that rely on them will provide man with access to growing amounts of knowledge of the kind that has previously been unavailable and could not be processed. The greatest breakthroughs that may result from the spread of such technologies are likely to occur in medicine and financial markets. I think it will take two or three years at the most for major changes to be seen clearly. Such changes are bound to impact all of us personally in a huge way.
Machine Learning Use Cases. Link
Machine Learning Infographics. Developed by Todd Jaquith and The Futurism Team