Welcome to SAIL!
At SAIL, we integrate Natural Language Processing (NLP), Human-Computer Interaction (HCI), and Statistics to examine language as a lens for analyzing communication breakdowns. Our mission is to address complex communication challenges by examining how language shapes social outcomes and designing AI technologies that promote meaningful and inclusive communication. Our research involves analyzing diverse, large-scale datasets, including decades of closed-caption data from major news outlets, police body-camera footage, online community interactions, and dialogues between neurodivergent and neurotypical individuals.
Research Areas
- Uncovering Language Patterns
- Understanding the Real-World Impact of Language
- Designing Human-Centered AI Language Technologies
By developing and applying NLP techniques, we analyze conversational trajectories and outcomes at scale to identify linguistic patterns linked to communication breakdowns. Our research spans contexts such as escalated police-citizen encounters, polarized media discourse, interactions between neurodivergent and neurotypical individuals , and responses from “bad actors” to online counterspeech.
To capture the real-world impact of language on society, we use statistical methods—including, Granger causality, and Structural Equation Modeling—to examine how communication patterns influence or predict outcomes such as public health trends, traffic stop results, and linguistic spillover effects from TV news broadcasts to social media discourse.
Guided by HCI principles and design thinking, we create language technologies that help users understand, anticipate, and prevent communication breakdowns. We work with experts in psychology, public health, law enforcement, and medical practitioners working with neurodivergent clients to address AI language limitations and ensure our tools are context-sensitive and user-focused.
Members
Current
Taufiq Daryanto
(PhD Student)
Alumni
Uma Gunturi
(MS Student)
Rohan Leeha
(MS Student)
Join Us!
We welcome researchers, students, and collaborators who are interested in joining or collaborating with the
Lab.
Please reach out at eugenia [at] vt [dot] edu.