2012-05-23

Co.Exist

IBM's Watson Could Help Mining Companies Stop Being So Destructive

The human mind has a hard time holding all the information needed to make informed decisions about future consequences. Not so with IBM’s Jeopardy-playing computer, which is now being used to give us a sense of what might happen when we dig giant, polluting holes in the ground.

Image by IBM.

Last year, IBM’s Watson artificial intelligence computer beat two Jeopardy champions in a much-publicized episode of the quiz show. It was a simultaneously awe-inspiring and terrifying party trick--proof that computers are finally starting to match human brainpower. But it was more than that. According to IBM, Watson’s debut heralded the era of cognitive systems (past eras include the tabulating era and the programmable computer era). In this era, computers like Watson can learn and predict complex consequences of actions better than humans. Now that this technology is starting to emerge, we need to figure out what to with it.

Enter the first-ever Watson academic case competition, which asked MBA students at The University of Rochester to come up with ways to solve societal and business challenges using Watson. "There’s a need for different customers to understand it and adopt it and understand the value of it. It’s equally important to build out a set of skills and talent pool so people can realize the value of the technology," explains Manoj Saxena, the general manager of IBM Watson Solutions. If the winners of the competition are anything to go by, there is plenty of value that can be derived from Watson in the future.

First place went to a team that explored how Watson could combine census data with weather data to better identify weather patterns and improve response times in a crisis. The third-place idea looked at how Watson could analyze unstructured information to cut down on travel security wait times without compromising safety. But our favorite was the second-place project, dubbed "Mining for Insights, Literally."

This project proposed using Watson to help energy companies take into account regulatory, safety, and environmental impact information in ways that are virtually impossible at the moment. Say a mining company has four different coal fields that it wants to evaluate. Today, that company will evaluate the impact of mining coal in these different areas based on a small set of data. But Watson can bring together newspaper reports, health reports, economic reports, safety reports, and all sorts of other pieces of unstructured data to create a composite understanding of relative risk and reward at each of the different coal fields.

Would every mining company follow up on the revelation that mining a specific field would lead to previously undetectable environmental consequences? Probably not. But just having the information available makes a difference, and the technology could be used "wherever things are harvested from the planet," according to Saxena. Think: logging, natural gas extraction, and gasoline drilling.

IBM is taking all the winning ideas into consideration for future projects, and it is already planning future case-study competitions at universities around the world.

Watson has already found its way into industry; a pilot program at WellPoint is using the technology to help doctors treat patients more effectively. Between 20% and 44% of the time, the first cancer diagnosis that a patient receives (i.e. what type of cancer they have) is wrong, says Saxena. "The reason is they’re not able to pull together all the medical information that’s available, all the recent developments from textbooks, journals, newspaper articles, medical studies and apply it at the point of care, because it’s humanly impossible for doctors to process so much information." Humanly impossible, but not impossible for Watson.

United Nations

Add New Comment

0 Comments