X

Clarkson prof in Potsdam leading team to evaluate DNA software

Posted 6/19/18

POTSDAM -- A Clarkson University professor is working with a $75,000 grant to lead a team in evaluating differences in the results given by a variety of software programs used to analyze mixed DNA …

This item is available in full to subscribers.

Please log in to continue

Log in

Clarkson prof in Potsdam leading team to evaluate DNA software

Posted

POTSDAM -- A Clarkson University professor is working with a $75,000 grant to lead a team in evaluating differences in the results given by a variety of software programs used to analyze mixed DNA samples.

Clarkson University Associate Computer Science Professor Jeanna Matthews is leading a Brown Institute Magic Grant-funded team that will systematically compare forensic DNA software which, taken together, are not providing consistently reliable evidence.

Prof. Matthews says DNA evidence in criminal and other cases of law will often have a serious and life-altering effect and should be examined for shortcomings in software analysis.

Matthews’ project, Decoding Difference in DNA Forensic Software, has been awarded a $75,000 grant, aimed at determining when, why, and by how the results of DNA mixture software programs differ. Real world cases have shown that differences can have a substantial impact on justice, including cases where one program incriminates a suspect while a different program excludes him. The project is expected to move beyond anecdotal examples to a systematic investigation leading to answers called for by the Presidential Council of Scientific Advisors on Science and Technology for independent testing of these DNA software programs.

“We will be comparing the results of probabilistic genotyping software systems,” Matthews said. “The results of these systems are treated as highly reliable evidence in court but they often disagree with each other and the producers of these systems aggressively resist third party review - even expert witness review under a protective order. This is one important example of black box programs being used to make big decisions about people’s lives. As citizens, we should expect rigorous explanations for how these decisions are being made and a feedback loop that identifies and corrects inevitable problems in software.”

The Brown Institute for Media Innovation, a collaboration between Stanford University’s School of Engineering and Columbia Journalism School, is awarding close to $1 million in funding for 12 projects as part of the 2018-19 Magic Grants.

Matthews’ team includes Nathan Adams, a DNA investigations specialist; Jessica Goldthwaite, a defense attorney at The Legal Aid Society; Dan Krane, a Biologist and Professor at Wright State University; Surya Mattu, a Pulitzer Prize-nominated Journalist with Gizmodo and ProPublica; and David Madigan, a statistician and professor at Columbia University.

Their research will systematically explore algorithmic accountability and transparency, and the role of complex software systems in the criminal justice system and elsewhere.