Thanks for telling me that hypergraphs are simplicial complexes... I reckon hypergraphs can be broken down as a list of subsets (from the powerset of nodes). This could also be viewed as a list of propositions, 1 proposition = 1 hyper-edge.
So, the hypergraph representation is pretty much equivalent to a set of propositions. In my new theory I'm still using the set-of-propositions as knowledge representation, albeit the propositions are mapped to vector space, such that they can be acted on by a deep neural net. I'm wondering if the hypergraph ≅ simplicial complex idea could lead to a drastically different kind of representation structure, unlike the set-of-proposition ones? 2) Even if you have probability distributions over the hypergraphs, that doesn't give you much leverage. The bottleneck of AGI is in the learning algorithm... I think we should focus on how to make *that* faster 😄 YKY ------------------------------------------- AGI Archives: https://www.listbox.com/member/archive/303/=now RSS Feed: https://www.listbox.com/member/archive/rss/303/21088071-f452e424 Modify Your Subscription: https://www.listbox.com/member/?member_id=21088071&id_secret=21088071-58d57657 Powered by Listbox: http://www.listbox.com
