VI.4 Human Speechome Project

Languages: English | Español | 汉语 | Deutsch

George Shaw

Phillip DeCamp

Deb Roy

This map by data visualization researchers George Shaw and Philip James DeCamp and cognitive scientist Deb Roy shows the experimental setup used to record the language development of Roy’s son at home. Approximately 10 hours of high-fidelity audio and video were recorded on a daily basis from birth to age three. The resulting corpus contains over 100,000 hours of multi-track recordings and constitutes the most comprehensive record of a child’s development made to date. New annotation, analysis, and visualization tools were developed to study and computationally model the early words uttered by the child by tracing back to the contexts in which those words were spoken to him by adults. As shown on the map, the results of this analysis comprise human movement traces, word birthplaces, and social networks. For most children, language development is steady, progressive, and, to a casual observer, effortless. But for some children—those with developmental delays due to biological or environmental causes—language is a major developmental hurdle. Understanding the regularities in home environments is essential to understanding mechanisms of language acquisition, causes of delay, and, ultimately, appropriate intervention procedures.


Roy, Deb, Philip DeCamp, Michael Fleischman, Peter Gorniak, Jethran Guinness, Rony Kubat, Michael Levit, Nikolaos Mavridis, Rupal Patel, Brandon Roy, Alexia Salata, and Stefanie Tellex. 2006. “The Human Speechome Project.” In Proceedings of the 28th Annual Cognitive Science Conference. 2059-2064.

Cognitive Machines Group. Cognitive Machines. “The Human Speechome Project.” Accessed September 21, 2011.

Shaw, George, Phillip Decamp, and Deb Roy. 2010. Human Speechome Project. Courtesy of Cognitive Machines Group, MIT Media Lab. In “6th Iteration (2009): Science Maps for Scholars,” Places & Spaces: Mapping Science, edited by Katy Börner and Elisha F. Hardy.

Acknowledgements: This exhibit is supported by the National Science Foundation under Grant No. IIS-0238261, CHE-0524661, IIS-0534909 and IIS-0715303, the James S. McDonnell Foundation; Thomson Reuters; the Cyberinfrastructure for Network Science Center, University Information Technology Services, and the School of Library and Information Science, all three at Indiana University. Some of the data used to generate the science maps is from the Web of Science by Thomson Reuters and Scopus by Elsevier. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the National Science Foundation.