Acquisition of semantic classes for adjectives from distributional evidence
Download
1 / 33

Acquisition of Semantic Classes for Adjectives from Distributional Evidence - PowerPoint PPT Presentation


  • 127 Views
  • Uploaded on

Acquisition of Semantic Classes for Adjectives from Distributional Evidence. Gemma Boleda Universitat Pompeu Fabra Barcelona. general picture. automatic classification of adjectives Catalan according to broad semantic characteristics clustering syntactic evidence. motivation.

loader
I am the owner, or an agent authorized to act on behalf of the owner, of the copyrighted work described.
capcha
Download Presentation

PowerPoint Slideshow about ' Acquisition of Semantic Classes for Adjectives from Distributional Evidence' - ninon


An Image/Link below is provided (as is) to download presentation

Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author.While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server.


- - - - - - - - - - - - - - - - - - - - - - - - - - E N D - - - - - - - - - - - - - - - - - - - - - - - - - -
Presentation Transcript
Acquisition of semantic classes for adjectives from distributional evidence

Acquisition of Semantic Classes for Adjectives from Distributional Evidence

Gemma Boleda

Universitat Pompeu Fabra

Barcelona


General picture
general picture Distributional Evidence

  • automatic classification of adjectives

    • Catalan

  • according to broad semantic characteristics

  • clustering

    • syntactic evidence


Motivation
motivation Distributional Evidence

  • Lexical Acquisition

    • infer properties of words

    • lexical bottleneck

      • both symbolic and statistical approaches

  • adjectives

    • determining NP reference

      • the French general

    • establishing properties of entities

      • this maimai is round and sweet


Motivation1
motivation Distributional Evidence

  • initial motivation: POS-tagging

    • 55% remaining ambiguity involves adjectives

      general francès:‘French general’ or ‘general French’?

  • observations

    • general tendencies in syntactic behaviour of adjectives

    • ... which correspond to broad semantic properties

  • generalisation: best at semantic level

    • low-level tasks (POS-tagging)

    • initial schema for lexical semantic representation


Approach
approach Distributional Evidence

  • no general, well established semantic classification

    • have to build and test ours!

  • clustering: unsupervised technique

    • groups objects according to feature distribution

    • does not depend on pre-classification

    • provides insight into the nature of the data

  • shallow approach to syntax: n-grams

    • limited syntactic distribution

    • local relationship to arguments

      => test feasibility


Outline

Boleda, Badia, Batlle (2004) Distributional Evidence

outline

  • adjective syntax and semantic classification

  • methodology

  • experiment 1

  • experiment 2

  • partial conclusions

  • outlook: rest of the thesis


Outline1
outline Distributional Evidence

  • adjective syntax and semantic classification

  • methodology

  • experiment 1

  • experiment 2

  • partial conclusions

  • outlook: rest of the thesis


Adjective syntax
adjective syntax Distributional Evidence

  • default function: noun modifier (92%)

    • right of the noun (default position: 72%)

    • some to the left (‘epithets’: 28%)

  • predicative uses unfrequent (7%), but significant


Two way classification
two-way classification Distributional Evidence

  • number of arguments

    • unary: pilota vermella ‘red ball’

    • binary: professor gelós de la Maria ‘teacher jealous of Maria’

  • ontological kind (Ontological Semantics)

    • basic:vermell ‘red’

    • object: malaltia pulmonar ‘pulmonary disease’ (=> lung)

    • event: propietat constitutiva ‘constitutive property’ (=> constitutes)


Ontological semantics
Ontological Semantics Distributional Evidence

  • coverage (ordinary cases)

  • machine tractability

  • explicit model of world: ontology

    • vermell => attribute::colour::red(x)

    • pulmonar => related-to::lung(x)

    • constitutiu => event::benef::constitute(x)

  • however: no commitment to particular framework


Rationale
rationale Distributional Evidence

  • observation: syntactic preferences correspond to semantic properties

  • hypothesis: we can use syntactic features to infer semantic classes


Outline2
outline Distributional Evidence

  • adjective syntax and semantic classification

  • methodology

  • experiment 1

  • experiment 2

  • conclusions and future work


Data and procedure
data and procedure Distributional Evidence

  • 2283 adjectives

    >50 times in 16 million word Catalan corpus

    • lemma and morphological info

  • cluster the whole set

    • perform different tasks on different subsets

      • tuning subset: choose features

      • Gold Standard: evaluation and analysis


  • Features and feature selection
    features and feature selection Distributional Evidence

    • features:

      • empirically chosen from blind distribution

      • double bigram, simplified POS-representation

    • tuning subset: 100 adjectives

      • choose features (distribution)


    Fig. A: Feature selection Distributional Evidence


    Analysis
    analysis Distributional Evidence

    • Gold Standard

      • 80 adjectives

      • annotated by 3 human judges, acceptable agreement (92 and 84%, .72 and .74 kappa)


    Outline3
    outline Distributional Evidence

    • adjective syntax and semantic classification

    • methodology

    • experiment 1

    • experiment 2

    • partial conclusions

    • outlook: rest of the thesis


    Experiment 1 unary binary
    experiment 1: unary / binary Distributional Evidence

    • final evaluation:10 features, raw percentage

      • clustering algorithm: k-means (cosine)

    • predictions:

      • binary adjectives cooccur with prepositions more frequently than unary ones

      • unary adjectives are more flexible


    Unary binary results

    unary (yellow) Distributional Evidence

    binary (red)

    Fig. B: Clusters vs. unary/binary

    unary / binary: results

    • agreement with Gold Standard:

      • 97%, kappa = 0.87

      • comparable to humans

    • features:


    Outline4
    outline Distributional Evidence

    • adjective syntax and semantic classification

    • methodology

    • experiment 1

    • experiment 2

    • partial conclusions

    • outlook: rest of the thesis


    Experiment 2 basic object event
    experiment 2: basic / object / event Distributional Evidence

    • final evaluation: 32 features, normalisation

      • clustering algorithm: k-means (cosine)

    • predictions:

      • basic adjectives are flexible, work as epithets, occur in predicative contexts, appear further from the noun

      • object adjectives appear rigidly after the noun

      • event adjectives tend to occur in predicative positions and do not act as epithets


    Basic object event results

    Fig C: Clusters vs. basic/event/object Distributional Evidence

    basic / object / event: results

    object (yellow)

    • agreement with Gold Standard:

      • 73%, kappa = 0.56

      • lower than humans

    • features:

    event (orange)

    basic (red)


    Basic object event error analysis

    Fig C: Clusters vs. basic/event/object Distributional Evidence

    Fig D: Clusters vs. unary/binary

    basic/object/event: error analysis

    • something has gone wrong!

      • characterisation of event adjectives

    basic adjectives with an object reading (polysemy)

    binary!

    unary event adjectives

    binary event adjectives


    Outline5
    outline Distributional Evidence

    • adjective syntax and semantic classification

    • methodology

    • experiment 1

    • experiment 2

    • partial conclusions

    • outlook: rest of the thesis


    Partial conclusions
    partial conclusions Distributional Evidence

    • overall, results seem to back up:

      • use of syntax-semantics interface for adjectives

      • linguistic predictions as to relevant features and differences across classes

      • shallow approach

    • unary / binary: piece of cake

      • few binary adjectives, but worth spotting (denote relationships)


    Partial conclusions1
    partial conclusions Distributional Evidence

    • basic / object / event: need reworking

      • object adjectives seem to be the most robust class

      • variation in basic adjectives (default class), polysemy

      • event adjectives: seem to behave much like basic adjectives with respect to features chosen => redefine class?


    Outline6
    outline Distributional Evidence

    • adjective syntax and semantic classification

    • methodology

    • experiment 1

    • experiment 2

    • partial conclusions

    • outlook: rest of the thesis


    Outlook rest of the thesis
    outlook: rest of the thesis Distributional Evidence

    • rethink classification

    • redefine features in light of results

    • integrate polysemy judgments into the experiment and analysis

    • perform experiments with other corpora


    Classification
    classification Distributional Evidence

    • what to do with event adjectives? cp.:

      • constitutiu ‘constitutive’ (“active”)

      • legible ‘readable’ (“passive”)

      • reproductor ‘reproducing’ (“active, habituality”)

    • yet another parameter: gradability

      • important for adjectives

      • should be easy to induce


    Better blind distribution or self defined features
    better blind distribution or self-defined features? Distributional Evidence

    • n-grams: sparseness, selection

    • other features?

      • account for different levels of description


    Polysemy
    polysemy Distributional Evidence

    • crucial aspect, explains much of results

    • difficult to integrate!

      • meaningless kappa values

    • alternatives?

      • clearer definition of polysemy within task

      • specific tests

      • other resources: dictionary?


    Other resources
    other resources Distributional Evidence

    • CUCWeb (208 million word)

      http://www.catedratelefonica.upf.es

    • test whether “more data is better data” (Mercer and Church 1993: 18-19)

      • advantages and challenges of Web corpora

    • current results: for verb subcategorisation experiment, results 12 points lower than using smaller, balanced, controled corpus


    Acquisition of semantic classes for adjectives from distributional evidence1

    Acquisition of Semantic Classes for Adjectives from Distributional Evidence

    Gemma Boleda

    Universitat Pompeu Fabra

    Barcelona


    ad