Re: The future of KR in retrospective

>> what is ontology building if not a giant classification exercise?

As far as know, predicting structure is still an active and recent area of
research.

I don't mean to suggest that this is a solved problem! Just that it is a
very closely related problem and that the NN aproach looks like it scales
much better in both building and inferencing and though it is error-prone,
it is robust to noise.

> > Explicit representation of knowledge is almost entirely absent in
> > connectionist systems.
> 
> Are you sure? Isn't for instance word embedding relying on sequence of words
> and as such take features from knowledge representation? Similarly, markov
> models rely on the probability of appearance of a given "token". The token
> can encode both sense and grammatical features.

I think so. Sure you have input and output tokens. But I think the "meaning"
(or the "semantics", or the "knowledge") is encoded in the mapping. That 
mapping is opaque and not really very good for answering "why?"

> > A child doesn't learn by being fed a bunch of facts and rules, a child
> > learns by example and a trial and error feedback loop.
>
> Again, this doesn't expel rules or dynamic programming. Somehow I connect
> logic to dynamic programming.

Sort of. Almost every rule governing language and behaviour and interaction
with the world is really very hard to figure out and explicitly state. Maybe
in some cases it is possible. But that's not what children do.

> Logic is the source of truth whereas connectionist approach provides a
> summary.

Interesting. My intuition is precisely the inverse!

Best wishes,

William Waites | wwaites@inf.ed.ac.uk
Laboratory for Foundations of Computer Science
School of Informatics, University of Edinburgh

-- 
The University of Edinburgh is a charitable body, registered in
Scotland, with registration number SC005336.

Received on Thursday, 27 June 2019 12:28:30 UTC