Commonsense 2017 - Thirteenth International Symposium on Commonsense Reasoning (Commonsense-2017)

Endowing computers with common sense is one of the major long-term goals of Artificial Intelligence research. Commonsense knowledge and reasoning are relevant for many applications of current interest. Examples include robot and human collaboration, transparent machine-learning systems that can explain their conclusions, social media and story understanding software, and dialogue systems. The recent resurgence of interest in commonsense reasoning reflects a wider societal reaction to current technological advances, such as the fact that “next year a law will come into operation in [EU] member states which gives everyone a right to an explanation of any decision affecting them that has been reached algorithmically” [Guardian newspaper, 14 April 2017].
Approaches to acquiring commonsense knowledge and performing commonsense reasoning may incorporate semantics-based representation and inference, machine learning, natural language processing, computer vision, and/or cognitive science. The symposium aims to encourage cross-fertilization between these and other techniques. The synthesis of multiple approaches is challenging, but could jump-start progress on many outstanding problems of commonsense reasoning.
We welcome a wide variety of submissions, including formal results, experimental results, demos, surveys, evaluations and comparisons of different approaches, and papers on methodological issues. While mathematical logic has traditionally been the primary lingua franca of the Symposium, we welcome all relevant and rigorous approaches to automating commonsense knowledge and reasoning.
Topics of interest include, but are not limited to:
* Semantics-based representations for specific commonsense domains, such as:
- Time, change, action, causality
- Commonsense physical and spatial reasoning
- Legal, biological, medical, and other scientific reasoning incorporating elements of common sense
- Mental states such as beliefs, intentions, and emotions
- Social activities and relationships
* Inference methods for commonsense reasoning, such as:
- Logic programming
- Probabilistic, heuristic, and approximate reasoning
- Nonmonotonic reasoning, belief revision and argumentation
- Abductive and inductive reasoning
- Textual Entailment
* Methods for creating commonsense knowledge bases, such as:
- Statistical and corpus-based techniques, including both traditional machine learning and deep learning
- Crowdsourcing
- Hand-crafting domain theories
- Hybrid methods
* Applications of commonsense reasoning, especially interdisciplinary research in the following areas:
- Natural language understanding (understanding discourse, question answering, semantic parsing)
- Image understanding
- Cognitive robotics and planning
- Web-based applications (search, internet of things)
- Support technologies (computer-aided instruction, home automation)
* Discussions of the science of commonsense reasoning research, including:
- Meta-theorems about commonsense theories and techniques
- Relation to other fields, such as philosophy, linguistics, cognitive psychology, game theory, and economics
- Challenge problem sets and benchmarking
By default accepted papers will be published shortly after the symposium in the CEUR Workshop Proceedings series. Authors may however opt out of publishing in CEUR, e.g. if they wish to publish their paper at another venue. All accepted papers will be made available on the commonsensereasoning.org website for the duration of the symposium. The publication of a journal special issue which would include selected and extended papers from Commonsense-2017 is currently under consideration; later versions of this call for papers will have more information on this subject.
Important Dates
- Submissions due: August 4, 2017
- Submission notification date: September 8, 2017
- Camera-ready versions due: September 22, 2017
- Symposium: November 6-8, 2017
Submissions
- Submissions will be made through EasyChair, at: https://easychair.org/conferences/?conf=commonsens...
- Papers are limited to 6 pages, prepared in IJCAI or AAAI format, using Letter or A4 sized paper, plus one additional page for references.
Review Process
Each paper will receive three blind peer reviews. Selection criteria include novelty, technical accuracy and rigor, significance and generalizability, relevance, and quality of writing.
Program Chairs
- Andrew S. Gordon, University of Southern California
- Rob Miller, University College London
- Gyorgy Turan, University of Illinois at Chicago and University of Szeged
Program Committee
- To be announced
Website: http://commonsensereasoning.org