When the late racehorse owner Merv Griffin created the television game show Jeopardy in 1964, he could not have imagined what recently transpired on the program when an IBM computer decisively defeated two of the best human players in Jeopardy’s history. One of the contestants had won 74 straight games hosted by Alex Trebek, another California racehorse owner and breeder.
Experts in artificial intelligence said that this feat by the computer, named Watson after IBM’s founder, would have been considered implausible just twenty years ago. The capabilities of artificial intelligence were limited, and the technology was very futuristic. It brought to mind a vision of Dr. Spock querying an intelligent and conversational computer on the television show Star Trek.
The New York Times wrote that when Watson competed against skilled human contestants on Jeopardy, “IBM researchers were tackling a game that requires not only encyclopedic recall, but also the ability to untangle convoluted and often opaque statements, a modicum of luck, and quick, strategic button pressing.”
Judging by the significant advances made in artificial intelligence in the past couple of decades, the increasing application of this technology in all walks of life is certain. The racing industry has intriguing possibilities in areas like handicapping, mating of bloodstock, veterinarian medicine, training, yearling selection, and racetrack marketing.
Artificial intelligence is being refined to assist humans in becoming better decision makers. For example, efforts are underway to develop a computer-based physician’s assistant that will answer doctor’s queries. A doctor can take the advice or disregard it just as he or she would if the analysis came from a human. While the accuracy of artificial intelligence will undoubtedly improve over time, computers, like humans, do make mistakes. Watson, the IBM computer, answered “What is Toronto?” to this question about U. S. cities: “Its largest airport is named for a World War II hero; its second largest for a World War II battle.”
Problems and questions that people encounter can broadly be classified into one of three types. First, routine ones have accepted rules-based answers. For instance, a simple test tells whether a mare is in foal. Second, some are in the form of puzzles that require a person (or a computer) to recognize patterns. To illustrate, diabetes have five or six symptoms that occur together. Finally, the most complex problems arise much less frequently and are approached through intuition and experimentation. The incidence of Mare Reproductive Loss Syndrome (MRLS) in Central Kentucky in 2001 defied a ready explanation and generated alternative hypotheses.
The value of artificial intelligence in racing and breeding should show up mostly in solving the higher-order problems. Mitigating on-track breakdowns, handicapping horse races, and pairing stallions with mares, all require pattern recognition, whereas perplexing maladies like laminitis and MRLS necessitate intuition and experimentation in prevention and treatment.
While computers and software are already being used in problem-solving tasks, accelerating improvements in artificial intelligence should make these present-day applications look primitive. To what extent artificial intelligence will eventually diminish the need for human decision making is unknowable. As Sherlock Holmes might have remarked to his physician sidekick, “as of now the answer is not elementary, my dear Watson.”
Copyright © 2011 Horse Racing Business
Originally published in the Blood-Horse. Used with permission.
Recent Comments