edu.illinois.cs.cogcomp.llm.comparators

Class LlmStringComparator

java.lang.Object

edu.illinois.cs.cogcomp.llm.comparators.LlmStringComparator

public class LlmStringComparator
extends Object

Field Summary

Fields

Modifier and Type	Field and Description
static String	DEFAULT_LLM_THRESHOLD
static String	LLM_THRESHOLD

Constructor Summary

Constructors

Constructor and Description
LlmStringComparator()
LlmStringComparator(ResourceManager rm_)
LlmStringComparator(ResourceManager rm_, edu.illinois.cs.cogcomp.mrcs.comparators.Comparator<String,edu.illinois.cs.cogcomp.mrcs.dataStructures.EntailmentResult> comparator_)

Method Summary

Modifier and Type	Method and Description
edu.illinois.cs.cogcomp.mrcs.dataStructures.Alignment<String>	alignNEStringArrays(String[] ne1_, String[] ne2_) determines best lexical level match of two arrays of Name Entity using NESim.
edu.illinois.cs.cogcomp.mrcs.dataStructures.Alignment<String>	alignSentences(String text_, String hyp_) convenience method to generate the alignment for two sentences specified as Strings.
edu.illinois.cs.cogcomp.mrcs.dataStructures.Alignment<String>	alignStringArrays(String[] textTokens_, String[] hypTokens_) generate a lexical alignment using the comparator specified at construction, and the
double	compareAnnotation(TextAnnotation source, TextAnnotation target) Convenience method to generate a scalar score for two input sentences passed as TextAnnotation.
double	compareStrings_(String source_, String target_) convenience method to generate a scalar score for two input sentences passed as Strings..
edu.illinois.cs.cogcomp.mrcs.dataStructures.EntailmentResult	determineEntailment(String[] textToks_, String[] hypToks_) determines best lexical level match of two arrays of String, each representing a tokenized sentence.
edu.illinois.cs.cogcomp.mrcs.dataStructures.EntailmentResult	determineEntailment(String text_, String hyp_) convenience method with sentences as Strings.
edu.illinois.cs.cogcomp.mrcs.dataStructures.EntailmentResult	scoreAlignment(edu.illinois.cs.cogcomp.mrcs.dataStructures.Alignment<String> alignment_) generate the result (score, label, etc.) for the given alignment using GreedyAlignmentScorer

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Field Detail

LLM_THRESHOLD

public static final String LLM_THRESHOLD

See Also:

Constant Field Values

DEFAULT_LLM_THRESHOLD

public static final String DEFAULT_LLM_THRESHOLD

See Also:

Constant Field Values

Constructor Detail

LlmStringComparator

public LlmStringComparator()
                    throws IOException

Throws:

IOException

LlmStringComparator

public LlmStringComparator(ResourceManager rm_)
                    throws IllegalArgumentException,
                           IOException

Throws:

IllegalArgumentException

IOException

LlmStringComparator

public LlmStringComparator(ResourceManager rm_,
                           edu.illinois.cs.cogcomp.mrcs.comparators.Comparator<String,edu.illinois.cs.cogcomp.mrcs.dataStructures.EntailmentResult> comparator_)
                    throws IOException

Throws:

IOException

Method Detail

determineEntailment

public edu.illinois.cs.cogcomp.mrcs.dataStructures.EntailmentResult determineEntailment(String text_,
                                                                                        String hyp_)
                                                                                 throws Exception

convenience method with sentences as Strings. Tokenizes using LBJTokenizer and runs Aligner, Filter specified at instantiation.

Parameters:

text_ - the Text (typically longer text span)

hyp_ - the Hypothesis (intent: is this true, given the Text?)

Returns:

EntailmentResult (contains a set of parameters describing the entailment decision)

Throws:

Exception

determineEntailment

public edu.illinois.cs.cogcomp.mrcs.dataStructures.EntailmentResult determineEntailment(String[] textToks_,
                                                                                        String[] hypToks_)
                                                                                 throws Exception

determines best lexical level match of two arrays of String, each representing a tokenized sentence. some tokens may be filtered depending on the WordListFilter params specified at construction.

Parameters:

textToks_ -

hypToks_ -

Returns:

Throws:

Exception

scoreAlignment

public edu.illinois.cs.cogcomp.mrcs.dataStructures.EntailmentResult scoreAlignment(edu.illinois.cs.cogcomp.mrcs.dataStructures.Alignment<String> alignment_)

generate the result (score, label, etc.) for the given alignment using GreedyAlignmentScorer

Parameters:

alignment_ -

Returns:

alignSentences

public edu.illinois.cs.cogcomp.mrcs.dataStructures.Alignment<String> alignSentences(String text_,
                                                                                    String hyp_)
                                                                             throws Exception

convenience method to generate the alignment for two sentences specified as Strings.

Parameters:

text_ -

hyp_ -

Returns:

Throws:

Exception

alignStringArrays

public edu.illinois.cs.cogcomp.mrcs.dataStructures.Alignment<String> alignStringArrays(String[] textTokens_,
                                                                                       String[] hypTokens_)
                                                                                throws Exception

generate a lexical alignment using the comparator specified at construction, and the

Parameters:

textTokens_ -

hypTokens_ -

Returns:

Throws:

Exception

alignNEStringArrays

public edu.illinois.cs.cogcomp.mrcs.dataStructures.Alignment<String> alignNEStringArrays(String[] ne1_,
                                                                                         String[] ne2_)
                                                                                  throws Exception

determines best lexical level match of two arrays of Name Entity using NESim.

Parameters:

ne1_ -

ne2_ -

Returns:

Throws:

Exception

compareAnnotation

public double compareAnnotation(TextAnnotation source,
                                TextAnnotation target)
                         throws Exception

Convenience method to generate a scalar score for two input sentences passed as TextAnnotation. The function is used only when NER is called and NER_CONLL view is generated. Compared with function determineEntailment(String text_, String hyp_) which generates score for two input sentences passed as String.

Parameters:

source_ -

target_ -

Returns:

similarity score (double) between 0 and 1.

Throws:

Exception

compareStrings_

public double compareStrings_(String source_,
                              String target_)
                       throws Exception

convenience method to generate a scalar score for two input sentences passed as Strings..

Parameters:

source_ -

target_ -

Returns:

similarity score (double) between 0 and 1.

Throws:

Exception