Symbolic / semantic / logical methods for building conversational agent - maybe ideas from Eva chatbot can help here?

[Alex]

Hi!

I asked this question in StackExchange https://cs.stackexchange.com/questions/82575/semantic-drt-methods-for-conversational-agents-chatbots-dialogue-systems-r and actually I don't hope to receive meaningful answer there. So - maybe ideas and experience from Eva chatbot can provide some answer (and references) to my question. So - in essence - today the natural language can be formalized enough - e.g. http://www.springer.com/gp/book/9783319504209 and https://academic.oup.com/jos and there are efforts to convert the formalized language into meaningful logics like http://virthost.vub.ac.be/lnaweb/ojs/index.php/LogiqueEtAnalyse/article/view/1873

So - if language can be formalized then - can we perceive the conversation as the logical inference process? Conversational agent is BDI or PSI agent that acts based on the logical inferences. Such chatbots should be built and they should be far more rational and human than agents built from the statistical methods or precompiled answers.

So - is Eva chatbot such logical/semantic conversational agent and is yes - then - how unique it is. Are the prior art references which I can read further or are there research trends which form the context of the Eva and from which I can receive encouragment for my own efforts.

I like logics and inference control (it should be realizable as BDI agent) and I can not believe that so much credit is given to the statistical and subsymbolic methods in this century.

[Nil Geisweiller]

Hi Alex,

embryonic experiments with an inference based chatbot were attempted a
year ago or so

https://www.youtube.com/watch?v=LduD7Et_cOs
https://www.youtube.com/watch?v=Gu6zrwAVD0k

the inferences themselves are genuine, however the inference control +
some pre and post processing are hardwired, for efficiency and
reproducibility reasons.

The code is here

https://github.com/opencog/opencog/tree/master/opencog/nlp/chatbot-psi

To bring that to the next level a few things need to be improved, like
relex2logic, inference control, temporal reasoning, perhaps more.

Nil

> --
> You received this message because you are subscribed to the Google
> Groups "opencog" group.
> To unsubscribe from this group and stop receiving emails from it, send
> an email to opencog+u...@googlegroups.com
> <mailto:opencog+u...@googlegroups.com>.
> To post to this group, send email to ope...@googlegroups.com
> <mailto:ope...@googlegroups.com>.
> Visit this group at https://groups.google.com/group/opencog.
> To view this discussion on the web visit
> https://groups.google.com/d/msgid/opencog/a68d6d4e-fe65-4557-93bc-5c606d44d359%40googlegroups.com
> <https://groups.google.com/d/msgid/opencog/a68d6d4e-fe65-4557-93bc-5c606d44d359%40googlegroups.com?utm_medium=email&utm_source=footer>.
> For more options, visit https://groups.google.com/d/optout.