Dynamic Semantic Web
miyabi:
Semantic Web of Knowledge, Semantic Web in 3 dimensions
When you create a hierarchy or classification, you thereby fix the context to a single object.
In the Scrapbox domain, it's called the "dead text" problem. Similarly, there would be a "dead IP" problem. GPT
Make the document more fine-grained rather than stand-alone.
Then we could curate more.
ken:
knowledge graph
LLMs move from static to dynamic graphs.
Network changes depending on context
Meaning-making becomes dynamic.
Information that "when A becomes possible, B becomes possible."
A number of elements are updated by the input that A is now possible, following the network
Dynamic Semantic Web
Discord:
sowawa> I sometimes think about the semantic web again when I think about AI too. Or rather, a reminder.
Ikumi Akatsuka> Dynamic Semantic Web
YukiJP> What is Semantic Web?
sowawa> Like a dynamic RAG.
nishio> Since humans can't see the entire web anyway, why not render only the "surroundings" and generate only the nearby areas while moving around?
Image like an auto-generated dungeon game.
miyabi> It's like a stream that curates and generates useful context within a certain perspective.
sowawa> That's a filter bubble,,
(Supplemental)
nishio: is @miyabi the first person to say this? I'd like to record my contribution as the person who started it.
miyabi: I think the statement itself is mine.
ryoar: as you said, I remember that @miyabi's comment led to Jamesy's expanding on it and connecting it to @kensuzuki's discussion of dynamic networks. I think it would be fun to express it graphically in 3D.
---
This page is auto-translated from /nishio/ダイナミックなセマンティックウェブ using DeepL. If you looks something interesting but the auto-translated English is not good enough to understand it, feel free to let me know at @nishio_en. I'm very happy to spread my thought to non-Japanese readers.