Constraint Grammar/Speed - Revision history

Unhammer: /* See also */

2022-03-23T12:56:15Z

Unhammer: /* See also */

2014-09-20T15:24:25Z

Unhammer: /* Speedy alternatives to CG */

2014-09-20T15:23:24Z

Speedy alternatives to CG

Unhammer: /* Speedy alternatives to CG */

2014-09-18T11:33:59Z

Speedy alternatives to CG

Unhammer: /* Speedy alternatives to CG */

2014-09-18T11:33:49Z

Speedy alternatives to CG

Unhammer: /* Speedy alternatives to CG */

2014-09-15T18:12:32Z

Speedy alternatives to CG

Unhammer: Unhammer moved page CG/Speed to Constraint Grammar/Speed

2014-09-15T18:09:35Z

Unhammer moved page CG/Speed to Constraint Grammar/Speed

Unhammer: Created page with "Tips on how to speed up your Constraint Grammar. ==Put popular rules first== If a cohort can be fully disambiguated early on, CG won't have to bother with that cohort later a..."

2014-09-15T18:09:03Z

Created page with "Tips on how to speed up your Constraint Grammar. ==Put popular rules first== If a cohort can be fully disambiguated early on, CG won't have to bother with that cohort later a..."

New page

Tips on how to speed up your Constraint Grammar.

==Put popular rules first==
If a cohort can be fully disambiguated early on, CG won't have to bother with that cohort later at all.

E.g. if you two rules that work on nouns, but one of them only matches if there's some rare verb in the context, try to put that rule after the first rule (as long as you still get the correct disambiguation!).

==Avoid slow rule types==

===Regexes===
Lots of regex matching ("foo.*bar"r) is typically slow.

===(*)===
A rule like

<code><nowiki>"<foo>" ADD (bar) (*) IF …</nowiki></code>

is much slower than

<code><nowiki>"<foo>" ADD (bar) ("<foo>") IF …</nowiki></code>

(vislcg3 might optimise that away in the future?)

==Speedy alternatives to CG==
* apertium-tagger is a lot faster than CG, but only supports the same types of rules that the tagger can learn (e.g. select or remove this bigram). (However, if you run apertium-tagger before the CG, your CG has less work to do.)

* fomacg (see [[User:David Nemeskey/GSOC progress 2013]]) is a project to create a finite-state version of CG, but is currently research stage.

[[Category:Constraint Grammar]]

@@ Line 33: / Line 33: @@
 * More tips on p.48 and on of http://www.hf.uio.no/iln/om/organisasjon/tekstlab/aktuelt/arrangementer/arkiv/CG08/CG3_Oslo.pdf#48
+<pre>
+– disambiguation gain: SELECT > REMOVE, frequent rules first
+  (also: avoid "ghost" checking rules), "heavy" sets first, POS
+  targets vs. word targets, target frequency (not used)
+– processing cost:
+  ● rules length (in number of contexts)
+  ● global > local contexts
+  ● NOT/C > simple check
+</pre>
 [[Category:Constraint Grammar]]

@@ Line 32: / Line 32: @@
 ==See also==
-* More tips on p.48 of http://www.hf.uio.no/iln/om/organisasjon/tekstlab/aktuelt/arrangementer/arkiv/CG08/CG3_Oslo.pdf#48
+* More tips on p.48 and on of http://www.hf.uio.no/iln/om/organisasjon/tekstlab/aktuelt/arrangementer/arkiv/CG08/CG3_Oslo.pdf#48
 [[Category:Constraint Grammar]]

@@ Line 30: / Line 30: @@
 * [[Constraint Grammar/Optimisation|future versions of vislcg3]] :-)
+==See also==
+* More tips on p.48 of http://www.hf.uio.no/iln/om/organisasjon/tekstlab/aktuelt/arrangementer/arkiv/CG08/CG3_Oslo.pdf#48
 [[Category:Constraint Grammar]]

← Older revision		Revision as of 11:33, 18 September 2014
Line 28:		Line 28:
	* fomacg (see [[User:David Nemeskey/GSOC progress 2013]]) is a project to create a finite-state version of CG, but is currently research stage.		* fomacg (see [[User:David Nemeskey/GSOC progress 2013]]) is a project to create a finite-state version of CG, but is currently research stage.

	* [[Constraint Grammar/~~Optimisations~~\|future versions of vislcg3]] :-)		* [[Constraint Grammar/Optimisation\|future versions of vislcg3]] :-)

← Older revision		Revision as of 11:33, 18 September 2014
Line 27:		Line 27:

	* fomacg (see [[User:David Nemeskey/GSOC progress 2013]]) is a project to create a finite-state version of CG, but is currently research stage.		* fomacg (see [[User:David Nemeskey/GSOC progress 2013]]) is a project to create a finite-state version of CG, but is currently research stage.

			* [[Constraint Grammar/Optimisations\|future versions of vislcg3]] :-)

@@ Line 24: / Line 24: @@
 ==Speedy alternatives to CG==
-* apertium-tagger is a lot faster than CG, but only supports the same types of rules that the tagger can learn (e.g. select or remove this bigram). (However, if you run apertium-tagger before the CG, your CG has less work to do.)
+* [[apertium-tagger]] is a lot faster than CG, but only supports the same types of rules that the tagger can learn (e.g. select or remove this bigram).
 * fomacg (see [[User:David Nemeskey/GSOC progress 2013]]) is a project to create a finite-state version of CG, but is currently research stage.

← Older revision	Revision as of 18:09, 15 September 2014
(No difference)