Mingtong Zhang

Carnegie Mellon University, University of Illinois Urbana-Champaign

Papers

Total Citations

H-Index

About

Mingtong Zhang is a pioneering researcher at the intersection of affective computing, 3D scene understanding, and robotic imitation learning. Their work bridges human-centered AI and autonomous systems, with a focus on making machines more perceptive of both human emotions and physical environments. Zhang’s most impactful contribution is the development of the *ElderReact* dataset (2019, 37 citations), a multimodal resource that enables automatic emotion recognition in aging adults—a critical step toward deploying empathetic intelligent agents and social robots in healthcare and assisted living. This work challenges the field’s bias toward younger populations and opens new avenues for gerontechnology. Beyond human affect, Zhang has advanced 3D scene understanding with *Beyond RGB: Scene-Property Synthesis with Neural Radiance Fields* (2023, 10 citations), introducing a generative approach to infer geometric and semantic properties from novel viewpoints, which is vital for robot perception. Most recently, their *Neural Dynamics Augmented Diffusion Policy* (2025) tackles data efficiency in robotic imitation learning, reducing the need for extensive demonstrations—a breakthrough for scalable robot training. With a growing citation record and contributions that span emotion recognition, neural rendering, and robotics, Zhang is shaping a future where AI systems are both emotionally aware and physically competent.