The Sixth Annual Conference on Advances in Cognitive Systems will take place from August 18 to 20, 2018, at Stanford University in the San Francisco Bay area. The meeting welcomes many types of research, including demonstrations of new capabilities, empirical studies of implemented systems, and formal analyses of complex tasks. The term "cognitive" refers to any artifact that thinks or reasons, whether or not it works the same way as humans.
The conference shares with early AI research an emphasis on high-level cognitive abilities, structured representations, and integrated systems. Thus, submissions on compositional inference, multi-step planning, and the acquisition of complex knowledge structures are especially relevant. However, the meeting is agnostic about the particular formalism to used to encode expertise, as well as whether expertise is constructed manually or learned from experience.
The purpose of the conference is to report progress toward the original goals of artificial intelligence and cognitive science, which aimed to explain the mind in computational terms and to reproduce a broad range of human abilities in computational artifacts. Some functional abilities that arise in this context include:
Some research communities already address such issues, including those dealing with cognitive architectures, commonsense reasoning, cognitive robotics, qualitative modeling, and many others. We encourage participation from anyone who is interested in computational approaches to complex cognition, human-level intelligence, and related topics.
We invite researchers to submit papers for presentation at the conference. Submissions should be no more than sixteen (16) pages in length. Accepted papers, allotted four pages for further elaboration, will appear in Advances in Cognitive Systems (ISSN 2324-8416), an electronic journal associated with the conference. Authors should follow the instructions provided at
which give detailed information about format and which provide Latex and Word templates. Any papers that diverge from this format or that exceed 16 pages will be returned without review. All papers must be submitted electronically at
Each submission should state clearly the problem it addresses, describe its response to this problem, specify the novel theoretical tenets, make explicit claims about the approach's capabilities or behavior, and provide convincing evidence to support these claims. Every paper should also discuss related efforts, examine limitations of the reported work, and outline plans for future research. The review form at
elaborates on these publication criteria. We encourage authors to examine the form before drafting their manuscripts to ensure their submissions address them.
Because the conference aims to encourage research toward a broader understanding of intelligence, its criteria for determining contributions differ from traditional ones. Progress can include demonstrating new functionality, integrating different facets of intelligence, presenting a novel approach to an established problem, explaining complex cognition in humans, and formally analyzing a difficult task. We also welcome submissions that pose new problems or describe testbeds that challenge existing approaches.
The conference aims to be as inclusive as possible to foster innovative research on the computational nature of intelligence. The FAQ page (http://www.cogsys.org/faq) attempts to clarify the scope of the event. Authors who have questions about whether their research is appropriate for the meeting should contact the Program Co-Chairs, Pat Langley (firstname.lastname@example.org) and Dongkyu Choi (email@example.com), for additional information.