Thursday, July 11, 2013

We're Entering the Era when Machines will "Learn" and "Think"

As we enter into the third major era of computing, Cognitive Computing represents a whole new approach to solving complex data and information analysis problems.
Eric W. Brown, Director of Watson Technologies, IBM

Ninety percent of the world's data was created in the last two years, and we are rapidly approaching a point where more than 80 percent of it is unstructured, including all of those documents, videos and audio files flying around the internet. In fact, one of the greatest challenges facing most businesses is how to make effective use of these enormous and growing volumes of data, something we call big data. In the medical field alone, the amount of information is doubling every five years, yet healthcare providers have precious little time to keep up with all of this information.

To address these challenges, IBM is working on the next era of computing systems, or the third major era of computing, which is called Cognitive Computing, or systems that can "learn" or "think."

As the third major era of computing, Cognitive Computing follows the first era, which consisted of tabulating machines and the second, the era of programmable computers. Cognitive systems represent a whole new approach to solving complex data and information analysis problems, in three ways: They use deep analytics for huge amounts of data; they have learning capabilities that allow the system to automatically learn and improve over time; and, they have natural interfaces between humans and computers.

IBM's Watson system is one of the first systems built as a Cognitive Computing system. Watson applies deep analytics to text and other unstructured data sources to extract meaning from the data, and applies inference and reasoning to solve complex problems. As a first step toward Cognitive Computing, Watson expands the boundaries of human cognition by providing humans with fast, efficient access to relevant knowledge trapped in huge volumes of unstructured data. This capability can be used for complex problem solving, such as helping health professionals to treat patients.

How does Watson work?

Watson provides significant value by using hundreds of analytics that apply natural language processing, information retrieval, text analysis, knowledge representation and reasoning, as well as machine learning, to understand complex problems, generate possible answers, and evaluate evidence from unstructured data. This processing is inspired by how we as humans solve problems. At the same time, it provides a look into the future where Cognitive Computing systems will perform tasks that previously, only humans could do, thus freeing up humans to apply our immense cognitive capabilities for significantly more complex problems.

Think about it: in the programmable systems era of computing, people had to think and solve problems the way a machine processes information. For the first time, Cognitive Computing systems will begin to move beyond being blunt instruments of numbers and words to representing events in the real world. Machines are beginning to understand our world as we do. They will learn, much in the way we do: through the senses of sight, smell, taste, hearing and touch.

In this new era of computing, our machines will teach us and be taught by us. They will know the world through diverse inputs -- digital and organic -- for the purpose of helping people see through complexity, overcome bias, keep up with the speed of information and make better decisions. Cognitive computing is the driving force behind this change.

About the author of this Post

Eric W. Brown, Director of Watson Technologies, IBM
Eric W. Brown is the Director and Principal Investigator for Watson Technologies at IBM's T. J. Watson Research Center. Eric earned his B.S. at the University of Vermont and M.S. and Ph.D. at the University of Massachusetts, all in Computer Science.Eric joined IBM in 1995 and has conducted research in information retrieval, document categorization, text analysis, question answering, bio-informatics and applications of automatic speech recognition. Since 2007 Eric has been a technical lead on the DeepQA project at IBM and the application of automatic, open domain question answering to build the Watson Question Answering system. The goal of Watson is to achieve human-level question answering performance. This goal was realized in February of 2011 when Watson beat Ken Jennings and Brad Rutter in a televised Jeopardy! exhibition match.

No comments:

Post a Comment