Speakers:

Yessica Herrera, Visiting Scholar, Northwestern Institute on Complex Systems
Talk Title: The Body Speaks: Visual Cues of Psychological Stress in Bodily Expressions
Abstract: Emotions shape body movement, yet the visual cues that signal psychological stress—distinct from other emotional states—remain poorly understood. Acute stress alters motor patterns and may produce subtle expressive markers. In this study, dancers performed creative improvisations under stress (induced via the Trier Social Stress Test) and in a control condition. Movements were video-recorded and rated by 25 non-expert observers (ages 18–23, all female) using qualitative parameters from Laban Movement Analysis—Weight, Flow, and Rhythm— alongside perceived stress levels. Our study shows that observers reliably identified stressed performances, associating stress with tense, less fluid, and rhythmically altered movement. These findings reveal nuanced visual cues of psychosocial stress in expressive motion and have implications for fields like dance, clinical assessment, and emotionally intelligent systems. In particular, this work supports the growing efforts to make robotic movement more meaningful to humans by applying insights from movement perception studies to improve the design of expressive and more likable robotic technologies.

Aakriti Kumar, Postdoctoral Fellow, Kellogg School of Management and the Northwestern Institute on Complex Systems
Talk Title: Large language models can provide expert-aligned judgments of empathic communication
Abstract: Large language models (LLMs) appear to excel at empathic communication in text-based conversations. But, how reliably can machines judge the nuances of empathic communication? We compare annotations by experts, crowd workers, and LLMs based on four empathic communication frameworks applied to four different datasets. Specifically, we investigate the inter-rater reliability of these three groups across 1,050 annotations of 200 conversations where one partner is sharing a problem, and the other is offering empathetic support. We find high but imperfect reliability between experts across most sub-components of empathic communication; inter-rater reliability between experts varies based on the clarity, complexity, and subjectivity of these sub-components. Furthermore, we find that LLMs approach expert level inter-rater reliability and surpass the inter-rater reliability between crowd workers and experts. Finally, we demonstrate that evaluating subjective annotation can be misleading with traditional classification metrics but clear and robust when evaluating with inter-rater reliability contextualized by an empirical ceiling.

Tingyu "Mark" Zhao, PhD Student, Department of Industrial Engineering and Management Sciences
Talk Title: Noise Filtering in Complex Networks
Abstract: Networks are powerful representations of complex systems, yet real-world network data are often corrupted by edge-level measurement inaccuracies, sampling biases, and incomplete observations, compromising analytical validity. Here, we introduce the Network Wiener Filter (NetWF), a principled method to filter edge noise that jointly leverages both network topology and explicit noise characterization, thereby enhancing downstream analyses and inferences. We demonstrate the efficacy of NetWF in two distinct settings: the Enron Corpus email network and the genetic interaction network of the yeast \textit{Saccharomyces cerevisiae}, noting promising results in both studies. Equipped with technologies such as NetWF, we advocate for error-aware network analysis, with the hope to usher in a new chapter of network science, one that embraces data imperfection as an inherent feature and learns to navigate it effectively.

Sign Up:

Sign up to present at a future Lightning Talk session. NICO Lightning Talks are open to graduate students, postdoctoral fellows, and visiting scholars.

Location:

In person: Chambers Hall, 600 Foster Street, Lower Level
Remote option: https://northwestern.zoom.us/j/95387714084
Passcode: NICO25

About the Speaker Series:

Wednesdays@NICO is a vibrant weekly seminar series focusing broadly on the topics of complex systems, data science and network science. It brings together attendees ranging from graduate students to senior faculty who span all of the schools across Northwestern, from applied math to sociology to biology and every discipline in-between. Please visit: https://bit.ly/WedatNICO for information on future speakers.

Related Info