[DL] CFP - Evaluation and Design of Generalist Systems (EDGeS): Challenges and methods for assessing the new generation of AI

[Amanda Hicks]

AAAI Spring Symposium Series - Palo Alto, CAEvaluation and Design of
Generalist Systems (EDGeS): Challenges and methods for assessing the new
generation of AI

With the advent of large domain-universal models, we are witnessing a trend
towards generalist AI systems, no longer restricted to narrow tasks.
Alongside this trend has been a resurgence of research in symbolic AI, to
support common sense reasoning, explanation, and learning with limited
training data. There are few extant strategies for assessing modern
generalist AI systems, symbolic AI systems, or combinations thereof;
assessing the next generation of AI will require novel tools, methods, and
benchmarks that address both reasoning and generalist systems, individually
and combined holistically.
Topics

In the interest of fostering discussion of methodologies for understanding
and assessing AI in the domains of reasoning and generativity, we are
accepting submissions on topics including (*not exhaustive)*:

   - Novel training protocols for achieving generalist performance in
   reasoning and generative tasks
   - Limitations of current approaches in AI/Machine learning
   - Novel methodologies to assess progress on increasingly general AI
   - Methods and tools for identification of vulnerabilities in modern
   reasoning systems
   - Quantifiable approaches to assessing ethical robustness of generalist
   AI systems
   - Relevant architectures involving neuro-symbolics, neural network-based
   foundation models, generative AI, common sense reasoning, statistical and
   relational AI
   - Architecture for systems capable of reasoned self-verification and
   ethical robustness
   - Description of novel systems that combine reasoning with generativity

Format

This format of the symposium will include invited talks and contributed
paper presentations by leading researchers and technical experts, as well
as panel discussions and group breakout sessions focusing on the
implementation of competency assessment in real autonomous systems.
Submissions

Participants will be invited to submit:

   - Full technical papers (6-8 pages)
   - Technical presentation (including abstract < 2 pages and biography of
   main speaker)
   - Position papers (4-6 pages)

Manuscripts must be submitted as PDFs via EasyChair:
https://easychair.org/conferences/?conf=sss23.
<https://easychair.org/conferences/?conf=sss23>
Please align with AAAI format:
https://www.aaai.org/Publications/Templates/AuthorKit22.zip.
<https://www.aaai.org/Publications/Templates/AuthorKit22.zip>
Important dates

   - Abstract submission: 16th of January 2023
   - Notification: 6th of February 2023
   - Registration: 27th of February 2023
   - Camera-ready submission: 3rd of March 2023

Organizing Committee

   - *Joscha Bach, Intel Labs, Committee Co-Chair*
   - *Amanda Hicks,* * Johns Hopkins University, Committee Co-Chair*
   - *Tetiana Grinberg,* Intel Labs
   - *John Beverley,* University at Buffalo
   - *Steven Rogers,* Air Force Research Laboratory
   - *Grant Passmore,* Imandra
   - *Ramin Hasani,* MIT CSAIL
   - *Casey Richardson,* S&P Global
   - *Richard Granger,* Dartmouth College
   - *Jascha Achterberg,* University of Cambridge
   - *Kristinn R. Thórisson,* Reykjavik University, IIIM
   - *Luc Steels,* Barcelona Supercomputing Center
   - *Yulia Sandamirskaya,* Neuromorphic Computing lead, Intel Labs

For more information: https://www.cognitive-ai.org/edges-23

-- 
Amanda Hicks, PhD
Applied Ontologist
Johns Hopkins University Applied Physics Laboratory
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.zih.tu-dresden.de/pipermail/dl/attachments/20221216/5212955a/attachment.htm>