Entries with tag university of california los angeles.

Researchers Use Crowdsourcing to Solve Computationally Challenging Problems

Scientists are turning to crowdsourcing to solve a particularly challenging research problem because computation alone has proven unable to do the job. The University of California, Los Angeles, scientists are trying to use X-ray crystallography, a method for identifying the atomic and molecular structure of a crystal, to gain phase information about crystalline material. This information would allow them to model a crystallized molecule, which could help determine the structure of molecules such as those in proteins, DNA, or other substances that could be useful for purposes such as developing medications or treatments. To crowdsource their problem, the UCLA scientists created an online computer game based on a genetic algorithm, CrowdPhase, in which players are given a set of random images of electron-density maps of the crystal reflection phases. The players then pick the two images that best fit the researchers’ various criteria from these sets of phases. The process continues with the players ultimately helping to select the fittest and final solution. The researchers used CrowdPhase (www.crowdphase.com) successfully for two different phasing-related puzzles. (EurekAlert)(International Union of Crystallography)(Biological Crystallography Online)

Google Glass Could Help Diagnose Public Health Problems

Researchers have developed a Google Glass application that can diagnose and track diseases. University of California, Los Angeles, researchers used  the camera in Google Glass, a wearable computer system, to take a picture of a rapid diagnostic test strip. Individuals place blood or fluid samples on a conventional test strip, which changes color to indicate the presence of a disease. The system can return the test results to the user in as few as eight seconds and can analyze the data to help track and analyze the spread of diseases. <<Linda: What analyzes the images to make the determination? You have to explain this part of the process.>> The UCLA team tested the approach with HIV and prostate-specific antigen tests. In tests, the system was accurate 99.6 percent of the time when the images were read. This system could be used in remote areas where conventional medical tools are not available or in settings such as disaster areas. The researchers published their work in ACS Nano. (EurekAlert)(University of California Los Angeles)(ACS Nano)

UCLA Researchers Make New Material for High-Performance Supercapacitors

University of California, Los Angeles scientists have created a material they say could be used to create powerful supercapacitors. The material, a synthesized form of niobium oxide, could be used to rapidly store and release energy. The technology could be used to rapidly charge many devices, including mobile electronics and industrial equipment. The scientists have developed electrodes using the material, but must undertake more research to create entire quick-charging devices. Cornell University and the Université Paul Sabatier researchers contributed to the work, which was published in the journal Nature Materials. (EurekAlert)(University of California Los Angeles)(Nature Materials)

Algorithm Predicts Twitter-Worthy News

A University of California, Los Angeles, doctoral candidate has developed an algorithm that could help individuals and organizations increase their visibility on Twitter. Roja Bandari used artificial intelligence to examine the frequency of links and relinks of news articles posted to Twitter and develop an algorithm that predicts with 84 percent accuracy whether an article will be popular. The algorithm judges an article to be popular if it expects more than 100 tweets will link to it and unpopular it predicts there will be fewer than 20 links. One factor that predicted an item’s popularity, Bandari found, was the organization credited with writing the article. Conventional print media outlets—such as the Los Angeles Times and Reuters—were not as popular as sites such as TechCrunch, Mashable, and the Huffington Post. Explaining how journalists or bloggers interested in increasing their social-media profile could use the findings, Bandari said, “If you’re a freelance reporter writing about tech, you don’t want to write for the Christian Science Monitor.” Bandari created the new algorithm with two researchers from Hewlett-Packard’s HP Labs while she worked at the company as an intern. She presented the work at the 2012 International Association for the Advancement of Artificial Intelligence Conference on Weblogs and Social Media. (PhysOrg)(University of California, Los Angeles)

Showing 4 results.