Seminar - Computational Biomarkers via Neurocomputation, Machine Learning and Feature Selection

Publication Date: 2017-09-05 12:06:48

Monday, September 11 2017 – 14h30
DEI Building (Polo II) - Room A5

Invited Speaker: Larry M. Manevitz,  Department of Computer Science and Head, Neurocomputation Laboratory, U. Haifa

Title: Computational Biomarkers via  Neurocomputation, Machine Learning and Feature Selection

Abstract: Biomarkers are objective biological signals that indicate certain physiological or cognitive states and diseases. Classically these are relatively simple indicators such as specific blood-tests for, e.g. diabetes.  Such "objective" indicators have been lacking for many diseases and cognitive states. However, recent advances in machine learning now allow us to identify complex biomarkers. We give some recent examples and indications of the computational methodology as time allows based on joint works with my students and collaborators.   Examples include diagnosis of Parkinson's disease from speech, identification of cognitive states from fMRI and EEG signals,  the discovery of a secondary declarative memory system in human adults ("fast mapping"),  the ability to "read minds" from fMRI signals such as classification of the valence of freely retrieved human memories.

Short Bio: Prof. Larry Manevitz is the Head of the Neurocomputation Laboratory at the U. Haifa for the last 18 years.  He has had many visiting positions including Otago University (Dunedin), Oxford University, U. Texas (Austin), Royal Holloway (London), NASA Ames Research Institute, Courant Institute (NYU),  Hebrew University,  U. Maryland,  Stanford University.   A native of Brooklyn, NY his doctoral degree was from Yale University under the supervision of Abraham Robinson.