Following the success of the well-attended First and Second Workshops on Cognitive Knowledge Acquisition and Applications, we are excited to continue this workshop series at IJCAI 2017 in Melbourne, Australia. This workshop focuses on disseminating work that bridges cognitive psychology and artificial intelligence in an informal setting that promotes lively discussion and community-building among the participants.
KEYNOTE SPEAKER: Mary-Anne Williams, University of Technology Sydney
Cognitive systems are able to learn and reason in a manner that facilitates their natural and fruitful interaction with humans. Ultimately, this interaction aims to extend and enhance human cognition, not by having cognitive systems operate as subsidiary workers that solve problems for humans, but by having cognitive systems act as expert assistants able to collaborate with humans and provide them with advice in a form compatible with how humans naturally process and understand information.
Knowledge acquisition is central to the design of such cognitive systems. Knowledge should be in a form that allows systems to explain their inferences and accept user feedback. At the same time, knowledge acquisition should exhibit characteristics akin to those of human learning, so that humans can relate to it and be able to interact with it as if it were a knowledgeable colleague. Thus, we mean “cognitive” in the workshop’s title to characterize both the form of knowledge and the process of its acquisition.
Knowledge acquisition is central to the design of such cognitive systems. Unlike the significant body of work on mining the web for facts or answers to specific questions (e.g., NELL, IBM’s Watson system for Jeopardy!), the workshop’s emphasis is on the acquisition of general knowledge that can be applied by a cognitive system in novel situations to elaborate what has been sensed with plausible and useful inferences. At the same time, the process of knowledge acquisition should exhibit characteristics akin to those of human learning, allowing the cognitive systems to explain their inferences and accept user feedback to improve their performance. We are interested in contributions that take a position and discuss the merits of simple and intuitive acquisition processes that could potentially err more (when typical humans would also err) versus the merits of acquisition processes that use computationally-heavy machinery to improve performance at the expense of psychological validity.
Since knowledge acquisition cannot proceed independently of other aspects of cognition, like perception, reasoning, and decision making, we also welcome contributions on other aspects of cognition, as long as they are directly tied to knowledge acquisition within a unified framework. We particularly encourage the demonstration of (prototype) cognitive systems that implement the proposed frameworks and discuss solutions to pragmatic concerns that had to be addressed.
We welcome ongoing and exciting preliminary work. Topics of interest include, but are not limited to:
- Formal frameworks for acquiring cognitive knowledge.
- Deep learning for acquiring cognitive knowledge.
- Principled evaluation of acquired cognitive knowledge.
- Psychologically-guided design of the acquisition process.
- Considerations related to scalability and parallelization.
- Active choice among available learning data/resources.
- Representation languages for cognitive knowledge.
- Static versus temporal/causal cognitive knowledge.
- Interaction of acquisition with perception and reasoning.
- Alternative acquisition methods (e.g., crowdsourcing).
- Acquisition from media other than text (e.g., video).
- Architecture and implementation of cognitive systems.
- Real-world applications that utilize cognitive knowledge.
As part of this third instantiation of the workshop, we particularly encourage work on the theme:
Intelligent Assistants: Explaining Inferences and Accepting User Feedback
May 27, 2017: Submission deadline
June 17, 2017: Acceptance notification
July 18, 2017: Final PDF file deadline
August 20, 2017: Workshop in Melbourne, Australia.
Papers must be formatted according to the IJCAI 2017 guidelines (http://ijcai-17.org/FormattingGuidelinesIJCAI-17.zip), and be at most 6 pages long, plus an additional bibliography page.
Submissions (in PDF) are accepted through EasyChair:
Loizos Michael, Open University of Cyprus
Erik T. Mueller, Capital One
David Buchanan, Elemental Cognition/Bridgewater Associates
Ernest Davis, New York University
James Fan, customerserviceai.com
Hannaneh Hajishirzi, University of Washington
Antonis Kakas, University of Cyprus
Zachary Kulis, Capital One
Joohyung Lee, Arizona State University
Rob Miller, University College London
Henry Minsky, Google/Nest Labs
J. William Murdock, IBM
Ravi Palla, Capital One
John Prager, IBM
Alessandra Russo, Imperial College London
Claudia Schulz, Imperial College London
Biplav Srivastava, IBM
Gyorgy Turan, University of Illinois at Chicago