Unsupervised Language Learning

In OpenCog and SingularityNET, we are working on Unsupervised Language Learning which has multiple goals:

  • Enable capturing structured information from unstructured texts
  • Help studying grammars of unknown languages and ontologies of their cultures
  • Simplify lives of computational linguists
  • Let robots getting up to speed with humans with natural language conversations

Out international team is now distributed across US, Hong Kong and multiple sites in Russia. We are now working on sort of “internal” alpha version which is expected to be available for wide public use in second half of this year.

7 Likes

What would be involved in this? I used to be hesitant, but I’m actually rapidly warming up to the possibilities.

(Making a Weaver Bot as we speak.:D)

1 Like

Give us more on that topic. For instants how can we can involed?
all the best from cologne

2 Likes

Hi, you can follow updates on github GitHub - singnet/language-learning: OpenCog Unsupervised Language Learning and join mailing list Redirecting to Google Groups. In particular, if you are concerned about English language and experienced in unsupervised parsing, you can take our latest challenge: Unsupervised Parser Challenge for Gutenberg Children corpus · Issue #220 · singnet/language-learning · GitHub
If you are interested in applying unsupervised language learning stack for other languages, you can try it as well. The longer description of our current work scope with resources is here: Unsupervised Language Learning (ULL) Design (DRAFT) - Google Docs

1 Like

Hi @LWFlouisa, see my earlier reply to @sablowski - besides that, regarding the bot development, you may find some more ideas on using the unsupervised language learning stack for speech production: ULL for speech generation and semantic comprehension - Google Docs

1 Like

Thanks! Bookmarked

1 Like