Reproduce “Neuro-Symbolic AI: Explainability, Challenges, and Future Trends” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Xin Zhang, Victor S. Sheng
Xin Zhang and Victor S. Sheng at Texas Tech University propose a novel two-factor classification framework for Neuro-Symbolic AI (NeSy AI) explainability, revealing that the vast majority (184 out of 191) of surveyed NeSy AI models still exhibit low to medium-low transparency, despite the paradigm's promise of enhanced interpretability.
Key insights
- The vast majority (184 out of 191) of surveyed NeSy AI studies exhibit low to medium-low explainability, challenging the assumption that NeSy AI inherently provides high transparency.
- There has been substantial growth in NeSy AI research, particularly from 2020 to 2023, with images and text being the most common input data types.
- Achieving unified representations, ensuring genuine transparency across hybrid components, and fostering sufficient cooperation between neural and symbolic paradigms remain critical challenges for the field.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.