A variety of AI tools from IBM, including cloud-based and desktop machine learning. Watson gained fame in 2011 as the supercomputer that beat two highly successful human Jeopardy winners. Named after Tom Watson, the founder of IBM, 200 million pages of content from encyclopedias, dictionaries, articles, books, news, literature, taxonomies and ontologies were mined for the contest. Using the semantic relationships from this data, Watson accepted Jeopardy questions in natural language and derived answers in real time within a few seconds.
Running SUSE Linux and Hadoop on a cluster of 90 POWER7 servers (nearly 3,000 processing cores), Watson was not online during the contest. All data were stored in its 16 terabytes of RAM for fast access. See
Deep Blue.
From Games to Practical Use
After Jeopardy, Watson was put to use making treatment decisions at Memorial Sloan-Kettering Cancer Center, as well as to improve the customer experience at Citibank. Today, Watson is used to solve myriad problems using machine learning and deep learning. See
IBM.