Jump to content

User:Mirceat/Compositional distributional semantics

From Wikipedia, the free encyclopedia

Compositional distributional semantics

Compositional distributional semantic models are an extension of distributional semantic models that characterize the semantics of entire phrases or sentences. This is achieved by composing the distributional representations of the words that sentences contain. Different approaches to composition have been explored, and are under discussion at established workshops such as SemEval.[1]

Simpler non-compositional models fail to capture the semantics of larger linguistic units as they ignore grammatical structure and logical words, which are crucial for their understanding.

[2] [3]

Approaches

[edit]
  • category theory
  • deep learning
  • Hadamard?

Applications

[edit]

Corpora

[edit]

See also

[edit]

References

[edit]
  1. ^ "SemEval-2014, Task 1".
  2. ^ "Deep Learning for Semantic Similarity" (PDF).
  3. ^ "Sentence Pair Scoring: Towards Unified Framework for Text Comprehension" (PDF).
[edit]
  • Semantic Text Similarity Dataset Hub [1]