Vasant G Honavar

Followers

1,430

Following

Co-authors

Mentions

217

Public Views

Dr. Vasant Honavar received his Ph.D. in 1990 from the University of Wisconsin Madison, specializing in Artificial Intelligence. In September 2013, Honavar joined the faculty of Penn State University where he currently serves as a Professor and Edward Frymoyer Chair of Information Science and Technology. He founded and directs the Center for Big Data Analytics and Discovery Informatics at Penn State University. He is also on the faculty of the Huck Institute of the Life Sciences, the Institute for Cyberscience, and the Bioinformatics and Genomics Graduate Program. In addition to research, graduate student supervision and teaching, he is responsible for developing new research and educational initiatives in Data Sciences and contributing to research initiatives in Life Sciences.

From 1990 to 2013, Honavar served on the faculty of Computer Science and of Bioinformatics and Computational Biology at Iowa State University (ISU). At ISU, he directed the Artificial Intelligence Research Laboratory (which he founded in 1990) and the Center for Computational Intelligence, Learning & Discovery (which he founded in 2005) and served as the associate chair (2001-2003) and chair (2003-2005) of the ISU Bioinformatics and Computational Biology Graduate Program, which he helped establish in 1999 with support from an Integrative Graduate Education and Research Training (IGERT) award.

During 2010-2013, Honavar served as a program director in the Information and Intelligent Systems Division of the Computer and Information Sciences and Engineering directorate of the National Science Foundation (NSF) during 2010-2013 while maintaining his research program at ISU. He led the Big Data Science and Engineering Program, established the NSF-OFR collaboration in Computational and Information Processing Approaches to and Infrastructure in support of, Financial Research and Analysis and Management, contributed to Smart and Connected Health, Information Integration and Informatics, Expeditions in Computing, Science of Learning Centers, Integrative Graduate Education and Research Training, Computing Research Infrastructure Programs.

Honavar has led research projects funded by NSF, NIH, and USDA that have resulted in foundational research contributions (documented in over 250 peer-reviewed publications) in Scalable approaches to building predictive models from large, distributed, semantically disparate data (big data); Constructing predictive models from sequence, image, text, multi-relational, graph-structured data; Eliciting causal information from multiple sources of observational and experimental data; Selective sharing of knowledge across disparate knowledge bases; Representing and reasoning about preferences; Composing complex services from components; and Applications in bioinformatics and computational biology (especially analysis and prediction of protein-protein, protein-DNA, and protein-RNA interactions and interfaces, B-cell and T-cell epitopes, post-translational modifications), Social Informatics, Health Informatics, Energy Informatics, Security Informatics, and related areas.

Honavar has received many awards and honors during his career including the National Science Foundation Director’s Award for Superior Accomplishment in 2013 for his leadership of the NSF Big Data Program, the National Science Foundation Director’s Award for Collaborative Integration in 2011, the Iowa Board of Regents Award for Faculty Excellence in 2007, the Iowa State University College of Liberal Arts and Sciences Award for Career Excellence in Research 2008, and the Iowa State University Margaret Ellen White Graduate Faculty Award in 2011. However, his proudest accomplishments are the 32 PhD students, 25 MS Students and several undergraduate researchers that he has worked with and mentored during his career.
Phone: 814-865-3141
Address: 301A Information Sciences and Technology
Pennsylvania State University
State College, PA 16802

less

InterestsView All (74)

Uploads

Papers by Vasant G Honavar

A Causal Lens for Peeking into Black Box Predictive Models: Predictive Model Interpretation via Causal Attribution

arXiv (Cornell University), Aug 1, 2020

With the increasing adoption of predictive models trained using machine learning across a wide ra... more With the increasing adoption of predictive models trained using machine learning across a wide range of high-stakes applications, e.g., health care, security, criminal justice, finance, and education, there is a growing need for effective techniques for explaining such models and their predictions. We aim to address this problem in settings where the predictive model is a black box; That is, we can only observe the response of the model to various inputs, but have no knowledge about the internal structure of the predictive model, its parameters, the objective function, and the algorithm used to optimize the model. We reduce the problem of interpreting a black box predictive model to that of estimating the causal effects of each of the model inputs on the model output, from observations of the model inputs and the corresponding outputs. We estimate the causal effects of model inputs on model output using variants of the Rubin Neyman potential outcomes framework for estimating causal effects from observational data. We show how the resulting causal attribution of responsibility for model output to the different model inputs can be used to interpret the predictive model and to explain its predictions. We present results of experiments that demonstrate the effectiveness of our approach to the interpretation of black box predictive models via causal attribution in the case of deep neural network models trained on one synthetic data set (where the input variables that impact the output variable are known by design) and two real-world data sets: Handwritten digit classification, and Parkinson's disease severity prediction. Because our approach does not require knowledge about the predictive model algorithm and is free of assumptions regarding the black box predictive model except that its input-output responses be observable, it can be applied, in principle, to any black box predictive model.

Vasant G Honavar

Uploads

Papers by Vasant G Honavar

Log In