Frank van Harmelen is presenting his keynote "Neuro-symbolic ≠ Neuro-semantic" claiming that none of the commonly used embeddings are capturing any semantics. "It's just wishful mnemonics." Obviously, we are doing something wrong in neuro-symbolic AI... ;-) "Semantics is predictable inference."
#GeNeSy2024 workshop at #eswc2024
GeNeSy 2024 website: https://sites.google.com/view/genesy2024/
Slides (from a similar talk): https://de.slideshare.net/slideshow/the-k-in-neurosymbolic-stands-for-knowledge/258584872
