bgoertzel

(this is mainly for Rodas) This is a follow-on to Issue #1302 After getting the Cogita-mod QA system specified there working, the next step will be integrate the above functionality...

Modify Cogita chatbot to be a simple QA system utilizing RelEx2Logic, the microplanner/SuREal, and Man Hin's fuzzy sub-hypergraph matcher

6

(This is intended mainly for Rodas.) Look at the code for Linas's previous OpenCog-based chatbot, https://github.com/opencog/opencog/tree/master/opencog/nlp/chatbot There is nothing too deep AI-wise here. However, it is of interest operationally, because...

write script using pattern-mined English=>Lojban rules to create a set of Lojbanic Atoms representing the logical interpretation of a given English sentence

Write Scheme script invoking Pattern Miner to learn rules mapping English link-grammar-link-combinations into lojbanic-Atom-combinations

Complete Lojban parser so it has near 100% coverage of Lojban language features

NLP comprehension code should assign sensible truth values to linguistic Atoms

5

When a new word-instance is added to the Atomspace, the corresponding WordNode and ConceptNode should get their truth value's count incremented by 1. This can be done using either SimpleTV...