Hi

guys have a question as well :

in which way the construction grammar related to dialogue generation ?? 

still dnt get it !! can anyone upload any to help me in that ??

> http://www.meta-guide.com/home/bibliography/google-scholar/construction-grammar-dialog-systems

It seems to me that “grammar parsing” is involved with the *de-construction* of natural language; so, how is natural language *re-constructed* or “generated” from parsed grammar ?!?  And, what are examples of a tools or systems that do this ?!?

I’m just finding it hard to believe that all “grammar parsing” dialog systems (grammar agents) depend on machine learning to generate natural language responses….

> http://www.meta-guide.com/home/bibliography/google-scholar/rule-learning-dialog-systems

Is it correct that “decision trees” are generated by “rule learning systems”?

> http://www.meta-guide.com/home/bibliography/google-scholar/learning-classifier-dialog-systems

And, are “rule learning systems” not otherwise known as “learning classifier systems”?

> http://books.google.com/books?id=r3OTcQo3C7QC

Reinforcement Learning for Adaptive Dialogue Systems: A Data-driven Methodology for Dialogue Management and Natural Language Generation (2011)

> http://www.meta-guide.com/home/bibliography/google-scholar/classifier-dialog-systems-2011

I looked at 365x academic papers from 2011 about dialog systems mentioning “classifier” (above).

> http://www.meta-guide.com/home/machine-learning/classifiers-in-dialog-systems

From those 2011 papers, I manually extracted some 122x classifier terms, or “types” of classifier (above).  I then scored those classifier types for “popularity” against the past 10 years of papers in Google Scholar (above).

> http://www.meta-guide.com/home/machine-learning/classification-algorithms-in-dialog-systems

Then I took the whole “Category:Classification_algorithms” from Wikipedia (circa 125x) and scored those for popularity against the past 10 years of papers in Google Scholar (above).

> http://www.meta-guide.com/home/machine-learning/best-dialog-system-classifiers

Looking at the top dozen from both lists, I created a new Meta Guide webpage, “Best Dialog System Classifiers” (above).

= = =

It is still not clear to me whether or not all or most grammar based (grammar parsing) dialog systems (grammar agents) necessarily require machine learning and/or classifiers in order to generate natural language responses ?!?

What is more clear to me is that I now suspect that the “learning classifiers” create the so-called “language model” that the dialog system uses to “map” replies….

For instance, three forms of feature may be incorporated into the classifier: part-of-speech (POS) tags, lexical features, and syntactic properties.

The classification function can be defined in several ways: a multi-nomial naive Bayes classifier, a n-gram based classifier, a classifier based on grammatical inference techniques, or a classifier based on neural networks.

Basically, classifiers may be statistical or rule-based.