"Microsoft Researchers Use Machine Learning Techniques to Help Advance HIV Vaccine Research"
Microsoft PressPass (02/23/05)
Microsoft Research is applying computer science algorithms to HIV vaccine development at the University of Perth in Australia and the University of Washington. The two universities are pursuing related approaches to developing an HIV vaccine based on specific identifying proteins called epitopes; the method addresses two of the most vexing problems about HIV vaccination--the tremendous diversity of the virus and the variation of immune system components in human populations. Scientists at the University of Perth have collected the largest set of HIV samples matched with different immune types and are using Microsoft spam-fighting algorithms to help find patterns correlating HIV epitope evolution and people's specific immune types. Eventually, the research is expected to yield vaccines that are tailored to specific populations based on circulating HIV strains and immune types. University of Perth professor Simon Mallal says the research could apply to other types of mutating viruses, such as hepatitis C. The University of Washington research is searching for ancestral HIV epitopes that will allow immune systems to identify and kill infected cells. Microsoft software that is used to condense digital video and audio files is being applied to libraries of supplemental epitopes included in the University of Washington vaccine, reducing the size of the vaccine models by half. Shortened models mean easier administration of treatment and lower manufacturing costs for the vaccine, says University of Washington microbiology professor James Mullins. Microsoft researcher David Heckerman says the computer science algorithms have surprisingly proven more accurate when dealing with biological systems than with the computer science problems they were originally created for.
Full Article: http://www.microsoft.com/presspass/features/2005/feb05/02-23HIVResearch.asp
So basically, Microsoft is using their anti-spam techniques with Machine Learning and applying it to HIV fighting. It would be interesting, that some day we would be sending in microbots into people and it would be acting as a viral blocker. Interesting........ I just wonder how we would update our software. :)
No comments:
Post a Comment