com.github.monnetproject.bliss.sparsemath

com.github.monnetproject.bliss.sparsemath from the Monnet Project's bliss project.

License	BSD 3-clause "New" or "Revised" License
Categories	Net
GroupId	com.github.monnetproject
ArtifactId	bliss.sparsemath
Last Version	1.18.4
Release Date	Apr 20, 2018
Type	jar
Description	com.github.monnetproject.bliss.sparsemath com.github.monnetproject.bliss.sparsemath from the Monnet Project's bliss project.
Project URL	https://github.com/monnetproject/bliss
Source Code Management	http://github.com/monnetproject/bliss/tree/master

Download bliss.sparsemath

Filename	Size
bliss.sparsemath-1.18.4.pom
bliss.sparsemath-1.18.4.jar	138 KB
bliss.sparsemath-1.18.4-sources.jar	77 KB
bliss.sparsemath-1.18.4-javadoc.jar	272 KB

How to add to project

Apache Maven

<!-- https://jarcasting.com/artifacts/com.github.monnetproject/bliss.sparsemath/ -->
<dependency>
    <groupId>com.github.monnetproject</groupId>
    <artifactId>bliss.sparsemath</artifactId>
    <version>1.18.4</version>
</dependency>

Gradle Groovy

// https://jarcasting.com/artifacts/com.github.monnetproject/bliss.sparsemath/
implementation 'com.github.monnetproject:bliss.sparsemath:1.18.4'

Gradle Kotlin

// https://jarcasting.com/artifacts/com.github.monnetproject/bliss.sparsemath/
implementation ("com.github.monnetproject:bliss.sparsemath:1.18.4")

Apache Buildr

'com.github.monnetproject:bliss.sparsemath:jar:1.18.4'

Apache Ivy

<dependency org="com.github.monnetproject" name="bliss.sparsemath" rev="1.18.4">
  <artifact name="bliss.sparsemath" type="jar" />
</dependency>

Groovy Grape

@Grapes(
@Grab(group='com.github.monnetproject', module='bliss.sparsemath', version='1.18.4')
)

Scala SBT

libraryDependencies += "com.github.monnetproject" % "bliss.sparsemath" % "1.18.4"

Leiningen

[com.github.monnetproject/bliss.sparsemath "1.18.4"]

Dependencies

compile (2)

Group / Artifact	Type	Version
junit : junit	jar	4.10
it.unimi.dsi : fastutil	jar	6.4.4

test (3)

Group / Artifact	Type	Version
org.apache.commons : commons-math	jar	2.2
org.apache.commons : commons-compress	jar	1.4.1
colt : colt	jar	1.2.0

Project Modules

There are no modules declared in this project.

Bilingual Similarity Suite (BLISS)

This package provides a set of tools for working with topic modelling and in particular in the cross-lingual case, and for application to machine translation. The following algorithms are implemented

• Latent Dirichlet Allocation
• Cross-Lingual Explicit Semantic Analysis

And the following are planned

• Kernel Explicit Semantic Analysis
• Latent Semantic Analysis
• Coupled Probabilistic Latent Semantic Analysis

Building

Translation Topics uses Maven to build, and can be simply installed with the following command

mvn install

Building a corpus

To build a corpus for this there are existing scripts that download the data from Wikipedia. These can be run with (for English to German)

./build-wikipedia-article.sh en de

Mate-finding trials

Mate-finding trials can be run with the following command, from the experiments sub-folder:

mvn exec:java -Dexec.mainClass=eu.monnetproject.bliss.experiments.MateFindingTrial 
       -Dexec.args="trainFile metricFactory W testFile"

Where W is the number of distinct tokens in the corpus and metricFactory is:

• eu.monnetproject.bliss.clesa.CLESA: For CL-ESA
• (More to come)

Language model adaptation

Language models can be trained with the following command (from the betalm folder)

mvn exec:java -Dexec.mainClass="betalm.compile" -Dexec.args="corpus.gz N wordMap W lmFile"

Where N is the order of the n-gram model and W the number of distinct tokens. To adapt to a specific document provide in addition to -Dexec.args the following flags

    -Dexec.args="-b METHOD -f file[.gz] ..." 

Where METHOD is one of

• COS_SIM
• NORMAL_COS_SIM
• KLD
• JACCARD
• DICE
• ROGERS_TANIMOTO
• DF_JACCARD
• DF_DICE
• WxWCLESA

Monnet Project

Versions

Version
1.18.4 Apr 20, 2018