Human Computation

The internet has given rise to human participation in computational systems, via social networking, crowdsourcing, collective intelligence, and other manifestations in which people contribute information processing that is central to system behaviors and outcomes. For example, the online verification service, reCAPTCHA provides a resource by which websites can determine if the inputs they are receiving come from a human or machine by asking the “user” to enter words seen in distorted text images. While reCAPTCHA may protect websites from mal-intentioned ‘bots’, it is also a resource that helps digitize the text of old books and newspapers by including failed OCR fragments among the examples of distorted text, and using human responses as the means to decode them.

humancomp

reCAPTCHA is just one of a growing body of examples of how machine and human effort, when thoughtfully organized, come together to create a distributed information processing system. The success of such Human Computation (HC) systems ultimately comes from coupling understanding of Computer Science sensibilities, especially in the design and implementation of computational systems, with understandings of human proclivities, capabilities, and behavior.

This expansive, interdisciplinary research space includes machine-mediated computation by groups of individuals (e.g., pipelined problem solving systems), aggregate analytic results by groups that result from individual information processing (e.g., prediction markets), distributed networks of human sensors (e.g., mash-ups), and many other varieties of information processing that derive from involvement of humans as computational agents in simple or complex systems.

Thus, HC draws participants from such CS fields as computational linguistics, computer graphics, computer theory, computer vision, databases, geospatial systems, human-computer interaction, information retrieval, Internet technology and e-commerce, semantic web, machine learning, multi-agent systems, programming languages, robotics, sensor networks, software engineering, and wireless computing. In addition to these CS fields, a variety of outside disciplines have more recently been brought to bear on HC research, including behavioral sciences, cognitive science, complexity science, cultural anthropology, economics, evolutionary biology, philosophy, and neuroscience. Knowledge from these outside areas informs participatory mechanisms, inspires architecture and algorithm design, and helps anticipate the ramifications of systems that engender thought, interaction, and behavior at the human-computer interface.

Despite the area’s relative youth, HC systems have been embedded in real-world settings where the combined efforts of people and computing are achieving outcomes of notable scope and significance, such as in disaster relief, medical research, planetary resilience, scientific discovery, and education. A national research thrust in human computation, as characterized by this CCC study, is timely due to its potential to advance Computer Science, bring a significant new form of “computational thinking” to the behavioral sciences, and address problems of societal and national significance. We seek to bring together world-class luminaries, thought leaders, and innovators to explore the past and prospective impact of human computation and to clearly delineate the research areas and activities that will directly lead to the most beneficial national and societal outcomes.