IBM’s WatsonHolland Davey, Jana Babouder-Matta, Chris Honeycutt, David Burr
Introduction • Super Computer developed by IBM Research • Named for IMB’s founder: Thomas J. Watson • Initially created for Jeopardy! Game Show • Dr. David Ferrucci leads the Watson project
Development• Search engines deliver thousands of results that match keywords• University’s have worked on a consistent question answering software for years• Programmed by 25 IBM scientists• Not connected to the internet
Jeopardy! Challenge • Set out to answer complex Jeopardy! Questions • Language is hard for computers because of “intended meaning” • During trials, it won 70% of practice games
Jeopardy! Challenge• February 2011: First computer to compete against humans in Jeopardy!• Defeated shows greatest two champions Ken Jennings and Brad Rutter
Could YOU beat Watson? http://www.youtube.com/watch?v=qpKoIfTu krA&feature=related
Purpose of Watson • Idea formed during IBM top executives brainstorming publicity stunts • Deep Blue– chess supercomputer defeated Garry Kasparov • IBM’s previous most advanced machine was slow and inaccurate • Overall goal: “to create a new generation of technology that can find answers from data more effectively than current search engines”
Technology • Question-answering Technology • Deep understanding of natural language. – Process and answer complex questions that have puns, irony, and/or riddles
Technology • Computer running Software called Deep QA • Runs on cluster of Power 750 computers – ten racks holding 90 servers, for a total of 2880 processor cores running DeepQA software and storage – Holds approximately one million books worth of information
Words from John Kelly, Director of IBM Research:“I want to create something that I cantake into every other retail industry, inthe transportation industry, you nameit.” “Any place where time is critical and you need to get advanced state-of-the-art information to the front decision-makers. Computers need to go from just being back-office calculating machines to improving the intelligence of people making decisions.”
Beyond Jeopardy • Medical Assistant: Memorial Sloan Kettering Cancer Center Partnership (March 2012) – Clinicians will “teach” Watson to review oncological case histories and come up with best diagnosis and treatment
Managerial Purposes • Financial Assistant: Working with Citi Bank – Help analyze customer needs – Process financial, economic, product, and client data – Help financial professionals make better decision • Could IBM Watson rival complex derivatives on the trade floor?
Advantages • Provides Services that revolve around the new, digital world • Gives immediate answers instead of search results • Healthcare uses – Diagnosis – Information Warehouse • Efficiency and Organization
Disadvantages • Cannot read PET and CT scans to identify tumors • Questions asked must be in text • Less human effort – Too reliant upon technology – Less personal interaction between doctors and patients • Not cognitive – Only manipulates symbols • Limits understanding and reasoning behind decisions
Competition• Thus far there are no other computers that are near the performance level of Watson• Microsoft and GE announced plans to create something similar to use in the healthcare industry – Aim to use analytics, high performance software technologies to deliver patient outcomes and clinical applications
Applications to SIT • New technology for a business’ Decision Making Processes – Organizational structure may shift based on allocation of decision making • Potentially eliminates the needs/advantages of Virtual Teams (especially in healthcare) • Changes Knowledge Management Processes
Discussion Question #1• What do you think about the future of this product, and the potential to have a super computer with voice recognition software? Could this lead to an iRobot/personal assistant?
Discussion Question #2• Could this eliminate the use for the education system all together?
Discussion Question #3• If you were a manager at a business, would you trust using Watson with decision making, or do you feel human decision making still is a more reliable source?