Re: Something on grammar

Perception pipeline is used in the literature, and some also talk about neural “pathways” through the brain. The term “pipeline” is used in the sense that information generally flows from lower level to higher levels of abstraction for perception. The different levels are coupled, e.g. semantic priming influences the syntactic choice of word class (part of speech), so some information flows from higher to lower levels of abstraction to guide lower level processing, and this is also needed for learning purposes.  

Cheers,
Dave

> On 22 Jan 2021, at 11:59, Christian Chiarcos <christian.chiarcos@web.de> wrote:
> 
> All of this is consistent with a pipelined approach to natural language understanding, where processing occurs concurrently at different stages along the pipeline. This avoids backtracking, but includes a mechanism to reprocess text from a problem word when a problem is detected.
> 
> I don't want to be petty-minded, but isn't "reprocess text from a problem word when a problem is detected" the very definition of backtracking ?  -- Just kidding ;) I agree that a pipelined approach with some concurrent processing is probably the most realistic thing a cognitive architecture can be made efficient. We don't really have a good term for that. Strictly speaking, that's not a pipeline (in the NLP sense, at least) but more a parallel laying of pipes -- if you will; the metaphor doesn't really fit either.
>  
> I find this exciting as there are many clues that point to the requirements for a functional model of NLU, NLG and language learning, and the challenge is to experiment with ideas for realising those requirements in a simple way. We will then be able to test how performance scales with the length of a sentence and other attributes.
> 
> Looking forward to explore that ;)
> 
> All the best,
> Christian

Dave Raggett <dsr@w3.org> http://www.w3.org/People/Raggett
W3C Data Activity Lead & W3C champion for the Web of things