Dynamic interpretation and manipulation of symbols in Neuro-Symbolic AI

Determine methods to enhance the dynamic interpretation and manipulation of symbolic representations within Neuro-Symbolic AI systems so that symbols can be contextually adjusted and operationalized during reasoning and learning without sacrificing robustness.

Background

The paper surveys recent advances in knowledge representation, including commonsense knowledge bases, event-based representations, and neuro-symbolic techniques for efficiency. Despite these developments, the authors explicitly identify unresolved questions about how neuro-symbolic systems can make symbolic representations dynamic—interpreting and manipulating symbols contextually during reasoning—rather than relying on static encodings.