Nowadays Artificial Intelligence is one of the most developed branches of computer science. In past few decades, computer scientists made some huge leaps in this field and its applications. One prominent example of Artificial Intelligence applications is Watson computer system, designed and developed by IBM. Named after Thomas Watson, the first President of IBM, this computer can answer the questions, asked in natural language. It took part in the quiz show, called “Jeopardy!” in 2011. It won the contest and received the first prize of one million US dollars.
Participating in the quiz show of “Jeopardy!” with a computer system was the idea behind the development of Watson. The effort, required to design the system, was huge. With the support from different universities, such as Rensselaer Polytechnic Institute, Massachusetts Institute of Technology, University of Texas, University of Massachusetts, University of Southern California and Carnegie Mellon University, the IBM engineers completed the development in mid 2000s. In 2007, Watson was able to give correct answers to 15% of the questions. After a further improvement the machine was able to outsmart human competitors in 2010.
IBM designed Watson as a question answering computer system. It has been described as an application of machine learning technology, knowledge representation and reasoning, information retrieval and natural language processing to the open domain question answering field. For scoring, analysis, evidence gathering and hypothesis generation, DeepQA technology, invented by IBM, is employed in this machine.
Watson is a highly optimized computer system, in which multiple Power7 processors are connected in parallel. The whole system comprises a bunch of 90 IBM Power 750 servers in addition to I/O and network devices and 16 terabytes of RAM. Every Power 750 server employs a Power7 eight core processor with 4 threads per core. It runs at 3.5GHz. The parallel processing capability of Power7 enhanced the computing capability of Watson, which is capable of processing 500 gigabytes per second. The total cost of Watson hardware is approximately 3 million US dollars. The Watson system could be easily assembled with the help of an experienced computer support provider.
Linux is used as an operating system by most of the supercomputers. Watson is no exception. SUSE Linux Enterprise Server runs very fast on Power7 processor, which makes it an ideal operating system for Watson system. Some of the salient features of SUSE Linux Enterprise Server, running on Watson system, are highly efficient memory management, enhanced system performance, advanced input/output capabilities, POSIX thread library,etc. Linux operating system could be customized for Watson with the help from a veteran tech support provider.
In 2011, IBM and Nuance Communications teamed up to develop the commercial version of Watson, which will use its advanced features to support medical diagnosis and treatment. It has also been suggested that Watson could be very much useful for legal work. For commercial applications this machine could be used with the assistance from an IBM support provider.
Some critics have argued that Watson cannot think in spite of its amazing capabilities and it is capable of manipulating symbols only. While it cannot be denied that Watson is not cognitive like human beings, the truth is that Watson is a high performance analytic system which is capable of learning, synthesizing and delivering outputs.