We invite contributions to our AAAI 2013 Fall Symposium titled “How Should Intelligence be Abstracted in AI Research: MDPs, Symbolic Representations, Artificial Neural Networks, or _____?”. Each subfield of AI has a different perspective on intelligence and unspoken assumptions about what is critical to recreate it computationally. To better understand such differences, we aim to bring together a diverse group of AI researchers interested in discussing and comparing how intelligence and processes that might create it are abstracted in various subfields. For example, such discussion may include honest examination of the strengths and weaknesses of different approaches, and what features of biological intelligence are crucial or unnecessary to include in algorithms.

We hope to encourage cross-pollination of ideas between researchers viewing intelligence in different ways (e.g. through the lens of MDPs or symbolic manipulation) and at different levels of abstraction (e.g. biologically-plausible neural simulations or restricted Boltzmann machines). One goal is to facilitate revising or creating new abstractions of intelligence and intelligence-generating processes. More information can be found here: http://www.cs.ucf.edu/~risi/AAAISymposium2013/

Contributions related to how intelligence can or should be abstracted algorithmically in artificial intelligence research are invited. Extended abstracts that summarize the results of a research program along these lines are most welcome, as are personal position papers or contributions describing speculative work or work in progress. Works bridging traditionally separate AI paradigms are encouraged. Participants should be open to inspiration from work and ideas in other subfields, and be willing to step outside their intellectual comfort zones.

Interested participants are encouraged to submit extended abstracts (no more than 2 pages), or full-length papers (up to 6 pages in AAAI format) in PDF format to sebastian.risi@cornell.edu. Accepted submissions will be published as citable papers in the AAAI technical report.

Areas of interest include but are not limited to:

- Different levels and types of knowledge representation and reasoning
- Abstractions of the following:
- Neural networks (e.g. deep learning networks, spiking ANNs, and plastic ANNs)
- Learning (e.g. machine learning and reinforcement learning)
- Biological development (e.g. generative and developmental systems, and developmental robotics)
- Evolutionary search (e.g. digital evolution and evolutionary algorithms)
- Biologically-inspired computation
- Evolutionary robotics
- Swarm intelligence
- Artificial life
- Philosophical arguments on characteristics of appropriate abstractions for AI

The symposium will be held Friday - Sunday, November 15-17 at the Westin Arlington Gateway in Arlington, Virginia (adjacent to Washington, DC).

** Keynote Speakers **
Andrew Ng via Skype (Stanford University, USA)
Georg Striedter (University of California, Irvine, USA)
Randall O'Reilly (University of Colorado, Boulder, USA)
Risto Miikkulainen (University of Texas at Austin, Austin, USA)
Gary Marcus (New York University, USA)
More TBD

** Schedule **
Full Paper/Extended Abstract Submission: May 24, 2013
Notification: June 21, 2013
Final Camera-ready Paper/Extended Abstract: September 12, 2013

We look forward to hearing from you. Thank you!

-- Sebastian Risi, Joel Lehman, Jeff Clune