Social Intelligence and Theory of Mind

CADE Project: Endowing an AI with Theory of Mind using probabilistic behavior languages

This work explores developing inverse behavior language - in particular inverse HAP developed by Charles River Analytics to do plan and goal recognition. We use a combination of PPL and visualization to allow experts to correct behavior models encoded.

Publications:

To come.

Team: Erica Kleinman, Stefany Arevalo Escobar

Alumni: Andy Bryant, Murtuza Shergadwala, Zhaoqing Teng

Partners: Charles River Analytics

Funders: ONR, CRA

Warmth and Competence

How to develop agents that engender trust? In this project we experimented with embedding gestures that can evoke warmth and competence in virtual embodied characters.

Publications:

Nguyen, T. H. D., Carstensdottir, E., Ngo, N., El-Nasr, M. S., Gray, M., Isaacowitz, D., & Desteno, D. (2015, August). Modeling warmth and competence in virtual characters. In International Conference on Intelligent Virtual Agents (pp. 167-180). Springer, Cham. [pdf]

Team: Elin Carstensdottir, Sharon Carnicke, Stacy Marsella

Funders: Northeastern University (tier 1 grant)

Endowing an AI with Theory of Mind using cognitive modeling

Publications:

Shergadwala, M. N., & Seif El-Nasr, M. (2021). Esports Agents with a Theory of Mind: Towards Better Engagement, Education, and Engineering. arXiv preprint arXiv:2103.04940. [PDF]

Team: Erica Kleinman, Murtuza Shergadwala, Zhaoqing Teng, Thanh Long X. Vu

Partners: Co-Design Lab at UC Berkeley

Funders: CITRIS

Learning Theory of Mind using Active Analysis and Crowd Sourcing

This work explores the use of crowd sourcing and active analysis as a technique to acquire a narrative model that can encode theory of mind and social intelligence.


Publications:

Feng, D., Sequeira, P., Carstensdottir, E., El-Nasr, M. S., & Marsella, S. (2018, December). Learning generative models of social interactions with humans-in-the-loop. In 2018 17th IEEE International Conference on Machine Learning and Applications (ICMLA) (pp. 509-516). IEEE. [PDF]
Feng, D., Carstensdottir, E., El-Nasr, M. S., & Marsella, S. (2019). Exploring improvisational approaches to social knowledge acquisition. [PDF]
Feng, D., Carstensdottir, E., Carnicke, S. M., El-Nasr, M. S., & Marsella, S. (2016, November). An active analysis and crowd sourced approach to social training. In International Conference on Interactive Digital Storytelling (pp. 156-167). Springer, Cham. [PDF]

Team: Dan Feng, Elin Carstensdottir, Sharon Carnicke, Stacy Marsella

Funders: NSF