Rada Mihalcea named Janice M. Jenkins Collegiate Professor of CSE

Mihalcea specializes in identifying the underlying emotions, beliefs, and values expressed in written text, and using that to draw inferences about the meaning of the text and certain characteristics of the writer.

Prof. Rada Mihalcea Enlarge
Prof. Rada Mihalcea

Prof. Rada Mihalcea, director of the University of Michigan Artificial Intelligence Lab, has been named the Janice M. Jenkins Collegiate Professor of Computer Science and Engineering for her distinguished record of research, teaching, and service.

Mihalcea performs research at the intersection of computer science and linguistics, studying natural language and human interaction. Her projects apply text mining to huge collections of written interactions – online comments, tweets, and other conversations – to identify properties such as deception and humor in text. She specializes in identifying the underlying emotions, beliefs, and values expressed in written text, and using that to draw inferences about the meaning of the text and certain characteristics of the writer.

Her Language and Information Technologies (LIT) research group uses these techniques to produce a number of important insights into how humans interact with one another, with computers, and even with news and information on the web. The applications of these studies have been broad, producing systems that can automatically detect fake news and methodologies to assist counselors working on substance abuse, medication adherence, and other behavior changes.

“My research focuses on natural language processing,” Mihalcea says of her work, “building computer programs that can do interesting things with language. I primarily concentrate on using language to understand people, and that includes computational social linguistics and conversational interactions and how language can be used to understand people and their behavior.”

Mihalcea has found that a great deal can be learned about a person’s beliefs, environment, and demographics based on the way they use language. In one recent project, Mihalcea and collaborators set out to predict a person’s emotions by examining the correlation between their tweets and environmental factors. By comparing factors like weather, news exposure, and social network emotion charge with the sentiments a user expressed on Twitter, the researchers determined that the cumulative sentiment expressed in the news that day is the second best predictor of user emotion. By combining all signals together, they were able to obtain an emotional predictive accuracy significantly better than using individual signals.

In another study, she and a team of researchers set out to build better language processing models that can account for the people behind the language, moving away from the “one-size-fits-all” strategy that is typically used in the development of natural language processing methods. They used word association exercises to pair different linguistic traits with over 176,000 participants’ demographics. Their results showed significant variations in the word associations made by groups based on gender and nationality. This finding has important implications primarily for minorities, whose language characteristics often “get lost” in the language spoken by the majority.

Rada Mihalcea's MIDAS lecture, "Learning Analytics with a Personal Touch"

“Among many other things, I learned a lot about how to collaborate effectively with other researchers during my PhD working with Rada. She really thrives in bringing together and leading groups with diverse backgrounds around a common interest or project. For everyone involved, I think that led to more wholistic perspectives about research problems, innovation, and more interesting discussions.”

-Dr. Steven Wilson, Postdoctoral Research Associate, University of Edinburgh
PhD 2019; advised by Prof. Mihalcea


Beyond building a deeper understanding of human emotions and their connection to language, Mihalcea has applied these techniques to several automated systems with bearing on major social issues.

In one notable instance, her research team designed an algorithm-based system that identifies the telltale linguistic cues in fake news stories, which could provide news aggregators and social media sites with a new weapon in the fight against misinformation.

“One very recent piece of research we’ve done in my group is the detection of fake news,” Mihalcea explained in a video for SAGE Online. “That’s where we use data that would potentially be used in social science studies to figure out why people are faking news and what are the characteristics of fake news. But we’ve done it completely computationally.”

The system was built using a collection of fake and legitimate news. It makes use of linguistic analysis, which analyzes quantifiable attributes like grammatical structure, word choice, punctuation and complexity. It works faster than humans and it can be used with a variety of different news types.


“She is always thinking about how to understand and help people, both in her research and in general. I’m grateful for having an advisor that cares so much about her students. If it weren’t for her I wouldn’t be at U-M.”

-Charlie Welch
PhD Candidate; advised by Prof. Mihalcea


Previous work by Mihalcea produced unique lie-detecting software that works by studying data from real world, high-stakes court cases. The system analyzed videos of the cases and drew conclusions from the speakers words and gestures. Unlike a polygraph, it didn’t require touching the subject or separately interrogating them. In experiments, it was up to 75 percent accurate in identifying who was being deceptive (as defined by trial outcomes), compared with humans’ scores of just above 50 percent.

Currently, Mihalcea is working on a U-M Precision Health Investigators Award to study how natural language processing can be used to enhance behavior counseling. The project seeks to develop computational methodologies that can assist counselors in their interventions and allow them to receive timely and cost-effective feedback. Her work is also being used to better understand changes in mental health, with particular emphasis on changes that occur during social distancing and self-isolation.


“In the five years that I’ve worked with her, Rada has advised me through various research projects, outreach work, and teaching. Her support has enabled me to succeed, and today, I consider her not only my advisor, but my mentor and friend as well.”

-Laura Burdick
PhD Candidate; advised by Prof. Mihalcea


In addition to her research, Mihalcea is known for her commitment to engaging more women in the field of computing. She has piloted a number of programs and courses to increase the pipeline and retention of women in engineering and computer science throughout her time with the department. Her efforts have introduced more women to the world of computing and encouraged them to pursue graduate studies and research careers in the field.

“We are trying to change the stereotypes,” she said in a Q&A with the U-M Business Engagement Center. “There is so much more to computer science than the general stereotypes out there. This field isn’t only for men and isn’t only about programming; there is so much we can do in terms of changing people’s lives for the better.”

In 2014, Mihalcea was pivotal in establishing and is the faculty champion for Girls Encoded, a program in CSE that hosts a number of outreach and social functions for female students. Together with CSE PhD student and Girls Encoded co-director Laura Burdick, Mihalcea has helped to design and organize numerous events to develop a pipeline for women to computer science.

In 2016, Mihalcea spearheaded the “Women in Computing” seminar series, which brings women to campus each year who have made exceptional contributions to the field of CS. These lectures are typically accompanied by an informal social event, allowing U-M students to meet and talk with these accomplished women.

In Fall of 2018, Mihalcea and Burdick launched an entry-level course designed specifically for students with no background experience in CS, specifically encouraging the enrollment of women and underrepresented minorities. In addition to this, Mihalcea and Burdick earned a Google grant to design a research mentorship program that gives undergraduate women in computer science valuable hands-on research experience.

Through these and other initiatives, Mihalcea has consistently demonstrated her commitment to making CS accessible and appealing to as many women as possible. To recognize her work in this area, Mihalcea has been recognized with both the U-M Center for the Education of Women (CEW+)’s 2018 Carol Hollenshead Award and the 2019 U-M Sarah Goddard Power Award.

Veronica Perez-Rosas, an Assistant Research Scientist working with Mihalcea in the LIT group, describes Mihalcea as an “amazing mentor, colleague, and friend.” Perez-Rosas earned her PhD from U. North Texas with Mihalcea as her advisor. She goes on to say:

“As her former PhD student, I cannot speak highly enough of her. She motivates her students and provides guidance and constructive feedback. She cares about students beyond their research activities and makes sure everyone feels part of the research group. She is very considerate and sympathetic. She is a great listener and motivates students to pursue their own interests and ideas. Rada is also very supportive of female students as she has led several outreach activities for their inclusion in Computer Science.

“Rada has made a great impact in my career. She is a role model of the researcher and professor I aspire to become. I admire her energy and passion to pursue new research problems. She has inspired me to pursue my own research ideas and encouraged me to take on new projects. I feel very fortunate that we are now colleagues and I can keep learning from her.”

Rada Mihalcea gives an introduction to Text Mining and computational methods for social science researchers.

Mihalcea received a PhD in Computer Science and Engineering from Southern Methodist University and a PhD in Linguistics from Oxford University. She joined the faculty at Michigan in 2013 after serving on the faculty at the Department of Computer Science at the University of North Texas. She has co-authored three books on graph-based natural language processing and text mining for social sciences, including a recent textbook titled Text Mining: A Guidebook for the Social Sciences. She is a Fellow of the Association for Computing Machinery (ACM), the recipient of an NSF CAREER Award, a Presidential Early Career Award for Scientists and Engineers, and an NSF INSPIRE (Integrated NSF Support Promoting Interdisciplinary Research and Education) Award. She served as general chair or program co-chair for the major conferences in her field. She currently serves as Vice President of the Association for Computational Linguistics.

About Janice M. Jenkins

In 1980, Janice Jenkins was the first woman hired as a faculty member in the U-M EECS Department. She didn’t start college until she had raised her five children, finishing her PhD at the age of 46.

At U-M, Jenkins made important contributions to automated arrhythmia analysis using advanced signal processing and computer techniques. She was director of the medical computing research laboratory (1981-2002), and of the digital design laboratory (1983-1998), an instructional laboratory on the design of microprocessor based systems that she initiated and taught.

In 1991, Jenkins received the U-M Sarah Goddard Power Award for her outstanding professional achievements and contributions to the education of women, and the NSF Faculty Award for Women in Science and Engineering. She is a Fellow of the Institute of Electrical and Electronics Engineers, the American Institute for Medical and Biological Engineering, and the American College of Cardiology. She has supported, mentored, and graduated 20 PhD students and eight MS thesis students.

Jenkins continued her research and advising work after her retirement in 2002, traveling back to U-M to work on an active NSF grant and supervise the research of two graduate and three undergraduate students. She also maintained a major role in an NIH/SBIR grant, and directed clinical studies at Loyola University Medical Center.