Dr. Katz’s research addresses cancer treatment communication, decision-making, and quality of care. His work aims to examine the dynamics of how precision medicine presents itself in the exam room via provider and patient communication and shared decision-making. Dr. Katz leads the Cancer Surveillance and Outcomes Research Team (CanSORT), an interdisciplinary research program centered at the University of Michigan and focused on population and intervention studies of the quality of care and outcomes of cancer detection and treatment in diverse populations. Dr. Katz and CanSORT have been collaborating with Surveillance, Epidemiology, and End Results (SEER) cancer registries since 2002 to study breast cancer treatment decision making at the population level. We obtain patient clinical and demographic information from SEER and combine this with surveys of patients and physicians to create comprehensive data sets that enable us to study testing and treatment trends and the challenges of individualizing treatments for breast cancer patients. In 2015 we added a new dimension to our research by partnering with evaluative testing firms to obtain tumor genomic and germline genetic test results for over 30,000 breast and ovarian cancer patients in the states of California and Georgia. We are also pursuing insurance claims data to assist with our analysis of physician network effects.
Dr. Abney has pursued research in natural language understanding and natural language learning, including information extraction, biomedical text processing, integrating text analysis into web search, robust and rapid partial parsing, stochastic grammars, spoken-language information systems, extraction of linguistic information from scanned page images, dependency-grammar induction for low-resource languages, and semisupervised learning.
The Language and Information Technologies (LIT) lab, directed by Rada Mihalcea, conducts research in natural language processing, information retrieval, and applied machine learning. The group specifically focuses on projects concerned with text semantics (word/text similarity, large semantic networks), behavior analysis (multilingual opinion analysis, multimodal models for deception detection, emotion recognition, alertness detection, stress/anxiety detection, analysis of counseling speech), big data for cross-cultural analysis (geotagging, understanding cross-cultural differences and worldview), educational applications (pedagogical search engines, automatic short answer grading, conversational technologies for student advising).
Several of the projects in the LIT lab are interdisciplinary, acknowledging the fact that language can be used to deepen our understanding in many different fields, such as psychology, sociology, history, and others. Some of the ongoing projects in the lab are collaborations with psychologists and sociologists, and target a rich modeling of human behavior through language analysis, seeking answers to questions such as “what are the core values of a culture?” and “are there differences in how different groups of people perceive the surrounding world?” The lab is also actively working on multimodal projects to track and understand human behavior, where language analysis is complemented with other channels such as facial expressions, gestures, and physiological signals.
Of interest, Prof. Mihalcea was quoted in a story about sexism and today’s virtual assistants such as Amazon’s Alexa, Apple’s Siri, and Microsoft’s Cortana; Refinery29.