Data Mining Finds Bad Drugs

University of Minnesota Duluth professor Ted Pedersen and University Minnesota Twin Cities professor Serguei Pakhomov have been awarded a three-year, $935,000 National Institutes of Health research grant to develop natural-language processing (NLP) techniques that search through medical records to quickly detect widespread adverse drug reactions. Pedersen says the goal of the project is to improve the quality of post-marketing surveillance for adverse drug reactions. He notes that although the Food and Drug Administration approves all drugs before making them available, people often take so many drug combinations that it is not possible to test every interaction. Additionally, pharmaceutical companies may not have conducted enough studies to identify possible adverse reactions, he says. Pedersen will use NLP to develop methods that can identify different statements that have similar underlying meanings in medical records to enable the quick identification of patients who are taking similar combinations of drugs and possibly suffering from adverse reactions.
University of Minnesota Duluth (01/21/09) Latto, Susan Beasy

An Algorithmic Game of Clue©

Duke University researchers have developed an algorithm capable of determining the best strategy for winning a game of CLUE©, a mathematical model that also could be used to help robotic mine sweepers find hidden explosives. Duke post-doctoral fellow Chenghui Cai says robotic sensors, like players in CLUE©, take information from their surroundings to help the robot maneuver around obstacles and find its target. "The key to success, both for the CLUE© player and the robots, is to not only take in the new information it discovers, but to use this new information to help guide its next move," Cai says. "This learning-adapting process continues until either the player has won the game, or the robot has found the mines." Artificial intelligence researchers call these situations "treasure hunt" problems, and have developed mathematical approaches to improving the chances of discovering the hidden treasure. Cai says the researchers found that players who implement the strategies based on the algorithm consistently outperform human players and other computer programs. Duke professor Silvia Ferrari, director of Duke's Laboratory for Intelligent Systems and Controls, says the algorithm is designed to maximize the ability to reach targets while minimizing the amount of movement.
Duke University News & Communications (01/27/09) Merritt, Richard

Is Technology Producing a Decline in Critical Thinking and Analysis?

University of California, Los Angeles (UCLA) professor Patricia Greenfield says that critical thinking and analysis skills decline the more people use technology, while visual skills improve. Greenfield, the director of UCLA's Children's Digital Media Center, analyzed more than 50 studies on learning and technology. She found that reading for pleasure improves thinking skills and engages the imagination in ways that visual media cannot. She says the increased use of technology in education will make evaluation methods that include visual media a better test for what students actually know, and will create students that are better at processing information. However, she cautions that most visual media does not allocate time for reflection, analysis, or imagination. "Studies show that reading develops imagination, induction, reflection, and critical thinking, as well as vocabulary," Greenfield says. "Students today have more visual literacy and less print literacy." Greenfield also analyzed a study that found that college students who watched "CNN Headline News" without the news crawl on the bottom of the screen remembered more facts from the broadcast that those who watched with the crawl. She says this study and others like it demonstrate that multi-tasking prevents people from obtaining a deeper understanding of information.
UCLA News (01/27/09) Wolpert, Stuart

Data Mining Promises to Dig Up New Drugs

European researchers have developed a robot called Eve that uses artificial intelligence, data mining, and knowledge discovery technology to analyze the results of the pharmacological experiments that it conducts. The robot can make informed decisions on how effective different chemical compounds will be at fighting diseases, potentially providing more effective treatments and a faster development process for medicines. Eve relates the chemical structure of different compounds to their pharmacological activity to learn which chemical compounds should be tested next. "Over time, Eve will learn to pick out the chemical compounds that are likely to be most effective against a certain target by analyzing data from past experiments and comparing chemical structures to their pharmacological properties," says Jozef Stefan Institute researcher Saso Dzeroski. Dzeroski says Eve should help scientists and pharmaceutical companies identify more effective compounds to treat diseases, and help them find drugs in a fraction of the time and cost of current methods. Dzeroski says Eve is the first robot-based computer system capable of originating its own experiments, physically performing them, interpreting the results, and repeating the cycle. He says that instead of choosing compounds for testing at random, Eve can pick compounds that are more likely to be effective.
ICT Results (02/02/09)