Detecting Sarcasm Using Different Forms Of Incongruity

What it is about

A one-page abstract of an invited talk summarising the author's thesis work. Sarcasm is treated as incongruity of two kinds: incongruity visible from the text and general background knowledge (captured with sentiment-based features, word embeddings and topic models), and incongruity that needs specific context (the author's earlier posts, or the conversation in dialogue). The abstract gives no data, method details or numbers; those are in the cited papers (Joshi, Sharma & Bhattacharyya 2015 and later work).

Ideas

Ideas: none for Mova beyond what is already built. prag/src/irony.rs implements the text-internal incongruity (explicit polarity flips, implicit positive verb + negative situation) after the full papers; this abstract adds only the split into text-internal vs context-dependent incongruity, which chapter 34 can cite as the organising idea. The context-dependent kind (the speaker's earlier stance, the conversation) is covered more concretely by ghosh-2017-type conversation-context work and by hee-2018-we-usually-don-t-like.

For Mova

34. Irony and figurative language §9–10 (computational irony; Mova's irony detector). Code: prag/src/irony.rs (already based on the full papers).

Cite

The paper itself, not this card. BibTeX from the library's record of the work. DOI 10.18653/v1/w17-5201

@inproceedings{joshi2017detecting,
  title = {{Detecting Sarcasm Using Different Forms Of Incongruity}},
  author = {Joshi, Aditya},
  booktitle = {Proceedings of the 8th Workshop on Computational Approaches to Subjectivity, Sentiment and Social Media Analysis},
  year = {2017},
  doi = {10.18653/v1/w17-5201},
  url = {https://doi.org/10.18653/v1/w17-5201}
}

This page in Markdown → https://md.mova.elamur.ai/free/en/docs/papers/joshi-2017-detecting-sarcasm-different-forms-incongruity/