“Our brains don’t need to tell our lungs to breathe or our hearts to pump blood,” writes Ravi Arimilli. “Unfortunately, computers require instructions for everything they do. But what if machines could analyze big data and determine what to do, based on the content of the data, without specific instructions?” [“Cognitive systems speculate on big data,” IBM Research, 26 February 2013] Fortunately for us, we no longer have to speculate about learning machines (i.e., artificial intelligence systems that can Sense, Think, Act, and Learn^®). They are already here. Systems like Enterra Solutions^® Cognitive Reasoning Platform™ (CRP) are capable of learning from large datasets and can even propose their own hypotheses to test concerning the data. Google certainly believes that cognitive systems are going to play an important role in the future of computing. On 26 January 2014, Google announced that it had made it largest ever European acquisition with the purchase of DeepMind Technologies for £400 million. Reports are that Google moved quickly to buy the company before Facebook went after the company. Late last year Facebook created an artificial intelligence team “to understand the emotions or behavior implied in posts to its site.” [“Google buys UK artificial intelligence start-up,” by Tim Bradshaw, Financial Times, 27 January 2014]

DeepMind has been very secretive about its activities. Bradshaw states it has been “operating in ‘stealth mode’.” The company’s website simply states: “DeepMind is a cutting edge artificial intelligence company. We combine the best techniques from machine learning and systems neuroscience to build powerful general-purpose learning algorithms. Founded by Demis Hassabis, Shane Legg and Mustafa Suleyman, the company is based in London and supported by some of the most iconic technology entrepreneurs and investors of the past decade. Our first commercial applications are in simulations, e-commerce and games.” DeepMind isn’t the only organization working on trying to blend AI and neuroscience. “In a paper … presented in December at the annual conference of the Neural Information Processing Systems Foundation, MIT researchers describe a new technique that expands the class of data sets whose structure can be efficiently deduced.” [“Machine learning branches out,” by Larry Hardesty, MIT Media Relations, 14 November 2013] The article continues:

“Not only that, but their technique naturally describes the data in a way that makes it much easier to work with. In the paper, they apply their technique to several sample data sets, including information about commercial airline flights. Using only flights’ scheduled and actual departure times, the algorithm can efficiently infer vital information about the propagation of flight delays through U.S. airports. It also identifies those airports where delays are most likely to have far-reaching repercussions, which makes it simpler to reason about the behavior of the network as a whole.”

In his article, Arimilli describes a patent that he and two colleagues were granted that “establishes a way for computer systems to analyze data in a whole new way, using ‘speculative’ population count (popcount) operations.” He explains that their technique goes beyond the process used by IBM’s famous Watson computer. He writes:

“[Our process] could reduce the number of instructions to analyze data down to a tenth of what standard popcount operations use today. The idea is based on a couple of principles:

• One is that while big data is, well, big, the output of popcount is small. Thus, speculating ahead of time on big data dramatically improves performance when the real time request occurs.

• Another is that even when the region of data to be analyzed is uncertain, the popcount can still generate a correct answer because of its cumulative nature. Basically, the speculative results accelerate obtaining the right answer.

Imagine Watson being able to speculate on what it’s being asked so that it combs through a domain of data in real time. And before the person who is asking the question even finishes speaking, the computer has already found the right answer. Speculative popcount technology is poised to be a foundational piece of this new era of cognitive systems because of how it efficiently analyzes any kind of data, across any domain. These systems will be able to cut through the big data to find the right data.”

Marcus Hutter, a Professor of Computer Science at Australian National University, heads another team that is working on the challenge of machine learning. He writes, “My research group has … successfully developed software that can learn to play Pac-Man from scratch.” [“To create a super-intelligent machine, start with an equation,” The Conversation, 28 November 2013] He notes that their system “is given no prior knowledge about these games; it is not even told the rules of the games! It starts as a blank canvas, and just by interacting with these environments, it figures out what is going on and learns how to behave well.” Mark van Rijmenam explains, “Machine learning is about creating algorithms and systems that can learn from the data they process and analyze. The more data is processed, the better the algorithm will become.” [“How Machine Learning Could Result In Great Applications for Your Business,” SmartData Collective, 11 January 2014] He continues:

“It is actually a science of getting computers to act without explicitly being programmed and is a branch of Artificial Intelligence (AI). AI is a scientific discipline to find patterns, extrapolate answers and make predictions using algorithms and computational techniques. Nowadays, Machine learning can be found in many applications, ranging from self-driving cars, to effective web search, facial recognition and speech recognition, so what could machine learning mean for your business? Machines learning can be a very powerful for your business, but for it to work it requires access to all available datasets. For Machine learning more data does mean better results, because new data will enable the computer program to teach and improve itself. It is a popular research topic as the potential of Machine learning is enormous. It could allow computers to become fundamentally different and more powerful than we are used to today, resulting in fascinating new applications.”

Google itself is no stranger to learning systems. “Three years ago, researchers at the secretive Google X lab in Mountain View, California, extracted some 10 million still images from YouTube videos and fed them into Google Brain — a network of 1,000 computers programmed to soak up the world much as a human toddler does. After three days looking for recurring patterns, Google Brain decided, all on its own, that there were certain repeating categories it could identify: human faces, human bodies and … cats.” [“Computer science: The learning machines,” by Nicola Jones, Nature, 8 January 2014] Jones continues:

“Deep learning itself is a revival of an even older idea for computing: neural networks. These systems, loosely inspired by the densely interconnected neurons of the brain, mimic human learning by changing the strength of simulated neural connections on the basis of experience. Google Brain, with about 1 million simulated neurons and 1 billion simulated connections, was ten times larger than any deep neural network before it. Project founder Andrew Ng, now director of the Artificial Intelligence Laboratory at Stanford University in California, has gone on to make deep-learning systems ten times larger again. Such advances make for exciting times in artificial intelligence (AI) — the often-frustrating attempt to get computers to think like humans.”

The goals for some cognitive computer systems are quite ambitious. I suspect, however, that tailored systems are going to dominate the business sector in the years ahead. Jones believes, “For computer scientists, deep learning could earn big profits.” Van Rijmenam adds, “The possibilities for machine learning are vast and over time it will result in smarter technology that has the accuracy of a computer and the adaptability of the most intelligent human beings.” If you haven’t started thinking about how a cognitive computing system could help advance your business, now would be a good time to start.