|Title||:||Cognitive Modeling of Information Search|
|Speaker||:||Saraschandra Karanam (Utrecht University, The Netherlands)|
|Details||:||Wed, 14 Dec, 2016 11:00 AM @ BSB 361|
|Abstract:||:||The dominant form of evaluating information retrieval systems that can be traced back to the Cranfield experiments is based on a standardized collection of queries, documents and relevance assessments. This methodology makes a number of simplistic assumptions regarding the user, task, relevance assessment and ignores the inherently interactive nature of the process of information search on the Internet. Not only is the process of information search interactive, it also involves a number of cognitive processes such as memory, attention, problem solving and decision making. Furthermore, a number of cognitive factors such as prior domain knowledge, age, level of experience with the Internet and spatial ability influence the above cognitive processes leading to a wide variation in the information search performance of users. The alternative to Cranfield-style evaluation: conducting in-situlaboratory experiments with real users is not practically feasible because of a number of factors such as non-availability of users and costs involved.
Modeling and simulation of information search provides a means to go beyond these limitations. Especially cognitive models, based on well-tested principles and theories of cognition, are capable of precisely characterizing complex cognitive activities such as the process of information search. This research talk describes two cognitive models of information search and demonstrates how they can be used to explain, simulate and predict individual differences in information search behaviour due to variations in two cognitive factors: age and prior domain knowledge. Support tools for search and navigation based on the predictions of cognitive models are also briefly discussed.
Bio: Saraschandra Karanam is a Postdoctoral researcher in the Interaction Technology group of the Department of Information and Computing Sciences, Utrecht University, Utrecht, The Netherlands. He has an interdisciplinary background: PhD in Computer Science with a specialization in Cognitive Science, Human-Computer Interaction and Usability & User Experience. His research focuses on cognitive factors underlying search and navigation, interactive information retrieval and cognitive modeling. His current research is directed towards adapting cognitive models of information search to simulate information search behaviour of older adults. Prior to this, he worked briefly in two industrial research labs: HP Labs and Xerox Research Centre-India, where he was busy developing applications that use crowdsourcing to optimise enterprise business processes. His work has so far led to the publication of 7 journal articles, 18 conference papers and 8 patents.