opennlp-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Jörn Kottmann (JIRA) <j...@apache.org>
Subject [jira] [Commented] (OPENNLP-199) Refactor the PerceptronTrainer class to address a couple of problems
Date Thu, 09 Jun 2011 21:07:58 GMT

    [ https://issues.apache.org/jira/browse/OPENNLP-199?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13046802#comment-13046802
] 

Jörn Kottmann commented on OPENNLP-199:
---------------------------------------

The error message is also a little confusing, because the expected value and actual value
are switched in the assertEquals which is done in this test. The test seems to be more accurate
on the build server than on your machine.

The test should ensure that the system is not broken, so it comes to the question when do
we consider the system as broken?

If a very minor code change changes the accuracy a little (maybe makes it more accurate) we
should not consider it as broken. So I believe we should use here an accuracy threshold which
must be exceeded. This way we fail the test if the accuracy goes down too much, but tolerate
any improvement.

What do you think?

> Refactor the PerceptronTrainer class to address a couple of problems
> --------------------------------------------------------------------
>
>                 Key: OPENNLP-199
>                 URL: https://issues.apache.org/jira/browse/OPENNLP-199
>             Project: OpenNLP
>          Issue Type: Improvement
>          Components: Maxent
>    Affects Versions: maxent-3.0.1-incubating
>            Reporter: Jörn Kottmann
>            Assignee: Jason Baldridge
>             Fix For: tools-1.5.2-incubating, maxent-3.0.2-incubating
>
>
> - Changed the update to be the actual perceptron update: when a label
>   that is not the gold label is chosen for an event, the parameters
>   associated with that label are decremented, and the parameters
>   associated with the gold label are incremented. I checked this
>   empirically on several datasets, and it works better than the
>   previous update (and it involves fewer updates).
> - stepsize is decreased by stepsize/1.05 on every iteration, ensuring
>   better stability toward the end of training. This is actually the
>   main reason that the training set accuracy obtained during parameter
>   update continued to be different from that computed when parameters
>   aren't updated. Now, the parameters don't jump as much in later
>   iterations, so things settle down and those two accuracies converge
>   if enough iterations are allowed.
> - Training set accuracy is computed once per iteration.
> - Training stops if the current training set accuracy changes less
>   than a given tolerance from the accuracies obtained in each of the
>   previous three iterations.
> - Averaging is done differently than before. Rather than doing an
>   immediate update, parameters are simply accumulated after iterations
>   (this makes the code much easier to understand/maintain). Also, not
>   every iteration is used, as this tends to give to much weight to the
>   final iterations, which don't actually differ that much from one
>   another. I tried a few things and found a simple method that works
>   well: sum the parameters from the first 20 iterations and then sum
>   parameters from any further iterations that are perfect squares (25,
>   36, 49, etc). This gets a good (diverse) sample of parameters for
>   averaging since the distance between subsequent parameter sets gets
>   larger as the number of iterations gets bigger.
> - Added prepositional phrase attachment dataset to
>   src/test/resources/data/ppa. This is done with permission from
>   Adwait Ratnarparkhi -- see the README for details. 
> - Created unit test to check perceptron training consistency, using
>   the prepositional phrase attachment data. It would be good to do the
>   same for maxent.
> - Added ListEventStream to make a stream out of List<Event>
> - Added some helper methods, e.g. maxIndex, to simplify the code in
>   the main algorithm.
> - The training stats aren't shown for every iteration. Now it is just
>   the first 10 and then every 10th iteration after that.
> - modelDistribution, params, evalParams and others are no longer class
>   variables. They have been pushed into the findParameters
>   method. Other variables could/should be made non-global too, but
>   leaving as is for now.

--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira

Mime
View raw message