Paul Ruvolo

pruvolo@cs.ucsd.edu Ph.D. Student Computer Science & Engineering University of California, San Diego

Research

I am fascinated by the prospect of creating machines that can thrive in chaotic social environments. I have been addressing this problem from a multidisciplinary perspective by collaborating with and working as a computer scientist and a developmental psychologist.

My current work is to use the formal mathematical framework of stochastic optimal control as a means of both programming robotic agents to interact with people and as a means of understanding the goals of human social actors (specifically mothers and infants engaged in face-to-face interaction).

I work in the Machine Perception Laboratory under Dr. Javier Movellan. I also collaborate closely with Daniel Messinger in the Psychology department at the University of Miami.

Publications and Presentations

• In Preparation / Submitted (e-mail if you are interested in a preprint)
• P. Ruvolo, D. Messinger, A. Fogel, and Javier R. Movellan. A Computational Approach to Identifying Intentional Behavior in Early Infancy
• G. Hofree, P. Ruvolo, M. Bartlett, P. Winkielman. Bridging the mechanical and human mind: Spontaneous mimicry of a humanoid robot
• P. Ruvolo, A. Moore, A. Chiba, M. O'Rourke, T. Wu, G. Littlewort, M. Bartlett, and J. Movellan. Detecting Lies and Mimicking Faces: Timing is the Key
• P. Ruvolo, J. Whitehill, and J. Movellan. Exploiting Structure in Crowdsourcing Tasks via Latent Factor Models
• P. Ruvolo, T. Wu, J. Movellan. Iterative Collocation and Control: An Algorithm for Approximately Optimal Control of Nonlinear Stochastic Systems in Continuous State, Action, and Time
• P. Ruvolo and J. Movellan. Goal-based Imitation in Continuous, State, Action and Time.
• Journal Publications
• Messinger D., Ekas N., Ruvolo P., Fogel A., Are you interested baby? Young infants exhibit stable patterns of attention during interaction. Infancy 2011.
• Messinger D., Ruvolo P., Ekas N., Fogel A., Applying Machine Learning to Infant Interaction: The Development's in the Details. Neural Networks 2010.
• Ruvolo P., Fasel I., Movellan J.,(2010) A learning approach to hierarchical feature selection and aggregation. Pattern Recognition Letters.
• T. Wu, N. Butko, P. Ruvolo, J. Whitehill, M. Bartlett, and J. Movellan. Multi-layer architectures for facial action unit recognition. IEEE Transactions on System, Man, and Cybernetics, To Appear.
• Conference and Workshop Publications
• Tingfan Wu, Juan Artigas, Whitney Mattson, Paul Ruvolo, Javier Movellan, and Daniel Messinger. Collecting a Developmental Dataset of Reaching Behavior: First Steps. IROS Workshop on Cognitive Neuroscience Robots 2011.
• Juan Artigas , Tingfan Wu, Paul Ruvolo, Whitney Matson, Daniel Messinger, Javier Movellan. Rethinking Motor Development and Learning. ICDL 2011.
• Ting-Fan Wu, Nicholas Butko, Paul Ruvolo, Jacob Whitehill, Marian Bartlett, Javier Movellan. Between-dataset AU Recognition Transfer. Face and Gesture Workshop 2011.
• Gwen Littlewort, Marian S. Bartlett, Jacob Whitehill, Ting-Fan Wu, Nicholas Butko, Paul Ruvolo, Javier Movellan. The motion in emotion -- a CERT baed approach to the FERA emotion challenge. Face and Gesture Workshop 2011.
• Ruvolo, P., Movellan J. (2010). Alternatives to Low-Level Synchrony Based Methods for Speech Detection. NIPS 2010.
• Finger, H., Ruvolo, P., Liu S., Movellan R. (2010). Approaches and Databases for Online Calibration of Binaural Sound Localization for Robotic Heads. IEEE International Conference on Intelligent Robots and Systems.
• Whitehill J., Ruvolo P., Wu T., Bergsma J., Movellan J. (2009) Whose Vote Should More: Optimal Integration of Labels from Labelers of Unknown Expertise. Neural Information Processing Systems (NIPS).
• Fasel I., Ruvolo P., Wu T., Movellan J. (2010) Infomax Control Approaches for Social Robotics. Neural Information Processing Systems (NIPS) Workshop on Probabilistic Approaches to Robotics Control.
• Wu T., Butko N., Ruvolo P., Bartlett M., Movellan J. (2009). Learning to Make Facial Expressions. International Conference on Development and Learning (ICDL).
• Ruvolo P., Fasel I., Movellan J. Optimization on a Budget: A Reinforcement Learning Approach. Neural Information Processing Systems (NIPS) (poster and poster spotlight).
• Ruvolo P., Fasel I., Movellan J. (2008). Tuning Optimizers for Time-Constrained Problems using Reinforcement Learning. Neural Information Processing Systems (NIPS) Workshop on Optimization for Machine Learning.
• Ruvolo P., Whitehill J., Virnes M., Movellan J. (2008). Building a More Effective Teaching Robot via Apprenticeship Learning. IEEE International Conference on Development and Learning (ICDL).
• Ruvolo P., Movellan J. (2008). Automatic Cry Detection in Early Childhood Education Settings. IEEE International Conference on Development and Learning (ICDL).
• Ruvolo P., Fasel I., Movellan J. (2008). Auditory mood detection for social and educational Robots. Internations Conference on Robotics and Automation.
• J.R. Movellan, F. Tanaka, I. Fasel, C. Taylor, P. Ruvolo, and M. Eckhardt, The RUBI project: a progress report, Proc. ACM/IEEE international conference on Human-robot Interaction, Arlington, VA, pp. 333-339, March 2007
• Workshop and Poster Presentations
• Ruvolo, P., Messinger, D., Fogel A., and Movellan J.R. (2011, May). Inferring Preferences from Behavior in Early Mother-Infant Interaction. Poster accepted to the Association for Psychological Science Annual Convention, Washington, D.C.
• Messinger, D., Ekas, N., Ruvolo, P., & Fogel, A. (2010, May). Young infants exhibit stable patterns of interest and affect during interaction. Poster accepted to the Association for Psychological Science Annual Convention, Boston, MA.
• Messinger, D., Ekas, N., Ruvolo, P., & Fogel, A. (2010, May). Early bi-directional influence: Competing agendas when infants gaze and mothers smile. Poster accepted to the Association for Psychological Science Annual Convention, Boston, MA.
• Fasel I., Taylor C., Ruvolo P., Movellan J. RUBIOS An Operating System for Exploring the Real Time Organization of Behavior. Joint Symposium on Neural Computation 2006.
• Ruvolo P., Fasel I., Movellan J. Supervised Learning of Human Speech Affect Using Viola-Jones Features. Joint Symposium on Neural Computation 2006.
• Demonstrations
• Ruvolo P., Bartlett M., Butko N., Lainscek C., Littlewort G., Whitehill J., Wu T., Movellan J. (2008). Machine Perception for Human Machine Interaction. Neural Information Prcessing Systems (NIPS).

Education

• UCSD - September 2005 - Present
  • NSF IGERT fellow in Vision and Learning in Humans and Machines (2005 - 2007).
• Harvey Mudd College, 2003
  • Major: Computer Science (BS)