With approximately 50 million scientific papers available in public databases– and a new one publishing nearly every 30 seconds – scientists cannot know about every relevant study when they are deciding where to take their research next.
A new tool in development by computational biologists at Baylor College of Medicine and analytics experts at IBM research and tested as a “proof-of-principle” may one day help researchers mine all public medical literature and formulate hypotheses that promise the greatest reward when pursuing new scientific studies.
Knowledge Integration Toolkit or KnIT. In a retrospective case study involving published data on p53, an important tumor suppressor protein, the team showed that this new resource called the Knowledge Integration Toolkit (KnIT) is an important first step in that direction, accurately predicting the existence of proteins that modify p53 – proteins that were subsequently found to do just that.
[Abstract]: http://dl.acm.org/citation.cfm?id=2623667
https://www.bcm.edu/news/research/automated-reasoning-hypothesis-generation
(Score: 2) by SlimmPickens on Wednesday August 27 2014, @07:30AM
One group make wicked spreadsheets, someone else uses a proprietary database and so on. There's sooo much data out there and yet you can't even have a little narrow AI look for correlation. Obviously this system is going to do that but who knows what "query the system" will end up meaning.
(Score: 2) by aristarchus on Wednesday August 27 2014, @08:07AM
formulate hypotheses that promise the greatest reward
Somehow, this does not seem to be conducive to good science. Is Monsanto involved? Or Big Pharma? Oh, IBM. Weren't they supposed to be dead? Quigley? Zombied, by now. No, really. So we have to do science for rewards now, kind of prostituting our superior intelligence to the highest bidder. And it's all automatic!
(Score: 2) by khallow on Wednesday August 27 2014, @11:14PM
Scientists make these choices all the time when they don't have infinite funding at their disposal. My concern would be that everyone would look at the same problems at the same time. A herd effect is rather harmful in a realm where incredible specialization in near unique fields yields better results.
The primary alternative is for your superior intelligence to be useless (since apparently your intelligence isn't superior enough that you don't have to care what the highest bidder wants). Doesn't sound like an improvement to me.
(Score: 2) by AnonTechie on Wednesday August 27 2014, @09:02AM
Here is some additional info about analyzing and processing BIG DATA.
Big Data presents scientists with unfolding opportunities, including, for instance, the possibility of discovering heterogeneous characteristics in the population leading to the development of personalized treatments and highly individualized services. But ever-expanding data sets introduce new challenges in terms of statistical analysis, bias sampling, computational costs, noise accumulation, spurious correlations, and measurement errors.
The era of Big Data – marked by a Big Bang-like explosion of information about everything from patterns of use of the World Wide Web to individual genomes – is being propelled by massive amounts of very high-dimensional or unstructured data, continuously produced and stored at a decreasing cost.
"In genomics we have seen a dramatic drop in price for whole genome sequencing," state Jianqing Fan and Han Liu, scientists at Princeton University, and Fang Han at Johns Hopkins. "This is also true in other areas such as social media analysis, biomedical imaging, high-frequency finance, analysis of surveillance videos and retail sales," they point out in a paper titled "Challenges of Big Data analysis" published in the Beijing-based journal National Science Review.
http://www.eurekalert.org/pub_releases/2014-08/scp-bbe082614.php [eurekalert.org]
[Paper]: http://nsr.oxfordjournals.org/content/1/2/293.full [oxfordjournals.org]
Albert Einstein - "Only two things are infinite, the universe and human stupidity, and I'm not sure about the former."
(Score: 2) by buswolley on Wednesday August 27 2014, @04:11PM
I was fine when AI started taking your jobs.
But now its trying to take over my job. Fuck that. : /
subicular junctures
(Score: 1, Funny) by Anonymous Coward on Wednesday August 27 2014, @04:14PM
Don't worry, you still can work as the computer's lab assistant.
For now.