Align-Disambiguate-Walk (ADW) Semantic Similarity System

An Align, Disambiguate, and Walk (ADW) Semantic Similarity System is a Semantic Similarity System that is based on an alignment and a random walk sense disambiguation algorithm.

• AKA: Alignment and Random-Walk-based Semantic Similarity System.
• Context:
  • It was initially developed and evaluated by Pilehvar et al., (2013) for solving 3 NLP tasks:
    • Semantic Textual Similarity,
    • Semantic Word Similarity,
    • Word Sense Coarsening Task.
  • It also is based on a semantic network topology.
• Example(s):
  • ADW-MF System,
  • …
• Counter-Example(s):
  • Bodenreider-Aubry-Burgun Semantic Similarity Measure,
  • XOA Semantic Similarity Measure,
  • SimRel-FunSim,
  • GraSM,
  • Generalized Lin's Semantic Ontology Term Similarity Measure,
  • Lin's Semantic Similarity Measure,
  • Jiang-Conrath Node-based Semantic Similarity Measure,
  • Resnik's Semantic Similarity Measure.
• See: Node-based Semantic Similarity Measure, Edge-based Semantic Similarity Measure, Topological Semantic Similarity Measure, Semantic Similarity Measure, Semantic Word Similarity Measure, Gene Semantic Similarity Measure, Semantic Relatedness Measure, Similarity Matrix, Semantic Word Similarity Benchmark Task, Semantic Similarity Benchmark Task.

---

---

References

2013

• (Pilehvar et al., 2013) ⇒ Mohammad Taher Pilehvar, David Jurgens, and Roberto Navigli. (2013). “Align, Disambiguate and Walk: A Unified Approach for Measuring Semantic Similarity.” In: Proceedings of the 51st Annual Meeting of the Association for Computational Linguistics (ACL 2013) Volume 1: Long Papers.
  • QUOTE: We present a unified approach to semantic similarity that operates at multiple levels, all the way from comparing word senses to comparing text documents. Our method leverages a common probabilistic representation over word senses in order to compare different types of linguistic data. This unified representation shows state-of-the-art performance on three tasks: semantic textual similarity, word similarity, and word sense coarsening.

    (...)

    Here we describe the configuration of our approach, which we have called Align, Disambiguate and Walk (ADW). The STS task uses human similarity judgments on an ordinal scale from 0 to 5. Therefore, in ADW we adopted a similar approach to generating similarity values to that adopted by other participating systems, whereby a supervised system is trained to combine multiple similarity judgments to produce a final rating consistent with the human annotators.

---