|Title||:||Machine Learning for Computer Aided Detection of Cancers|
|Speaker||:||Balaji Krishnapuram (Siemens Medical Services, USA)|
|Details||:||Fri, 9 Jan, 2015 4:00 PM @ BSB 361|
|Abstract:||:||Advances in technologies such as MRI & CT scans and digital pathology
have enormously increased the volume of electronic data available for
early detection of cancers and other abnormalities. However the
radiologist/pathologist is now overloaded with enormous quantities of
information although they do not have more time to read images. As a
result, the recent exponential growth in electronic information only
provides marginal improvements in diagnostic accuracy for patients in
practice. Computer Aided Detection (CAD) algorithms help by
identifying and clearly drawing attention to abnormalities.
Radiologists assisted by CAD technology detect abnormalities at a much
earlier stage, reduce the probability of erroneously overlooking such
life threatening lesions, and thus significantly improved patient
outcomes as compared to radiologists without access to such
In this talk we will describe the technical/mathematical challenges for machine learning (ML) algorithms for CAD. We will investigate a series of modeling ideas & novel algorithms that illustrate the benefit of approaching problems from first principles. Using this example we will illustrate transferable ideas that will help the audience better understand how ML research is conducted during the development of commercially successful products. At the end of the talk, undergraduate & graduate students previously exposed to an introductory machine learning class should be able to understand how they can further develop their skills to be more successful in commercial settings while developing ML products.
Bio: Balaji Krishnapuram leads the Health Services Innovation Center for Siemens Medical Solutions USA, a group responsible for launching new products/businesses based on disruptive technological innovation. He led teams that launched 7 commercially successful products using Machine learning over the last 10 years. He organized over 15 international conferences/workshops, serving as the General Chair of ACM SIGKDD 2010 and ACM SIGKDD 2016. He edited a book, authored over 25 patents and published over 50 articles in the leading journals and research conferences in the areas of machine learning, information extraction, personalized medicine etc. He obtained his B Tech from IIT Kharagpur in 1999, and PhD from Duke University in 2004.