While progress has certainly been made in smart/intelligent machines R&D over the past century, we believe that this can and should be sped up by a sharper focus on theory and methodology: The rate of scientific progress rests on the development of appropriate methodologies, but good methodologies can only be conceived in light of good theories. Improvements in both are unavoidable if we want to see the field of AI move confidently forward. Clear definitions are the hallmark of good scientific work: Not its subject matter - intelligence - that is, after all what we are seeking to understand (given definitions can only be working definitions at best), but rather, what questions we seek to answer, how we fashion our methods for answering them, and how and which conclusions may be drawn from the results. Any methodology is considered good if it is fruitful, but without boldness, may continue at a snail’s pace.
Theoretical focus
- Answers scientific and methodological questions about cognition and intelligence
- Describes theoretical considerations regarding architecture and/or methodology
- Explains how it will affect/ improve AI R&D
Applied focus
- Describes the use of a particular approach to system/software architecture and/or methodology
- May involve integration of technologies, data sources, functions, and/or methods
- Explains how it will affect/ improve AI development and/or deployment
Page limit: 40 (excluding references - extensions possible if called for by the paper’s topic needs)