AI + Communication Disorders 

Our AI+Communication Disorders project focuses on developing advanced AI systems to support individuals with speech and language challenges. Current efforts include building models to assist people with aphasia by interpreting co-speech gestures, as well as creating diagnostic tools that integrate motor and linguistic features to better understand developmental language disorder (DLD) in children. Leveraging multimodal fusion techniques and generative AI, this research aims to enhance diagnosis, intervention, and overall communication outcomes.

AI + Mental Health 

Our AI+Mental Health project aims to advance early detection and intervention by developing models that integrate multiple modalities of patient data. These include visual cues such as facial expressions and body gestures, acoustic features like tone of voice, and linguistic information from speech and written notes. By fusing these diverse signals, our models can help diagnose mental health symptoms and predict risk levels, enabling timely and personalized interventions. Ultimately, this research seeks to support clinicians with objective, data-driven insights for improved mental health care.

AI + Healthcare (Data Synthesis)

This project focuses on developing generative models that capture the complex structures and constraints of healthcare data to create high-quality synthetic datasets. These synthetic datasets provide a powerful alternative to real patient data, helping protect privacy, reduce bias, and support reproducible research. Ultimately, this work enables safer data sharing and more equitable AI-driven healthcare solutions. In addition, synthetic data can be used to augment limited datasets, improving model training in low-resource scenarios. 

AI + Journalism 

This project explores how AI can expand access to reliable news by generating reader-friendly journalistic articles tailored to diverse audiences, particularly those who may lack access to local news. By integrating an AI chatbot, the system engages directly with users to understand their interests and needs, producing personalized articles that foster inclusivity and engagement. At the same time, the project places strong emphasis on safeguarding against disinformation, with mechanisms to detect and prevent the spread of inaccurate content that could emerge from gen-AI models.

AI + Social Media 

This project investigates how people interact and engage across online platforms by developing graph-based models that capture the dynamics of digital communities. We study diverse domains, ranging from social media marketing, such as understanding influencer impact and audience engagement, to mental health communities where peer interactions play a critical role in support and risk detection. To achieve this, we collect and analyze datasets from platforms including YouTube, Instagram, TikTok, and Reddit, enabling us to model complex network behaviors at scale.

AI + Criminology 

This project focuses on developing AI systems that can predict and help prevent crimes by analyzing regional patterns and correlations with geographic, demographic, and other contextual factors. By leveraging graph-based models, we aim to capture complex relationships within communities and across regions to improve the accuracy of crime prediction. A key focus of this project is combating online human trafficking, where we design models to detect suspicious patterns and signals in digital spaces. Ultimately, this work seeks to provide data-driven insights that support law enforcement and policymakers in promoting safer communities both offline and online.