In natural language processing, textual entailment (TE), also known as natural language inference (NLI), is a directional relation between text fragments. The relation holds whenever the truth of one text fragment follows from another text.

Definition

edit

In the TE framework, the entailing and entailed texts are termed text (t) and hypothesis (h), respectively. Textual entailment is not the same as pure logical entailment – it has a more relaxed definition: "t entails h" (th) if, typically, a human reading t would infer that h is most likely true.[1] (Alternatively: th if and only if, typically, a human reading t would be justified in inferring the proposition expressed by h from the proposition expressed by t.[2]) The relation is directional because even if "t entails h", the reverse "h entails t" is much less certain.[3][4]

Determining whether this relationship holds is an informal task, one which sometimes overlaps with the formal tasks of formal semantics (satisfying a strict condition will usually imply satisfaction of a less strict conditioned); additionally, textual entailment partially subsumes word entailment.

Examples

edit

Textual entailment can be illustrated with examples of three different relations:[5]

An example of a positive TE (text entails hypothesis) is:

  • text: If you help the needy, God will reward you.
hypothesis: Giving money to a poor man has good consequences.

An example of a negative TE (text contradicts hypothesis) is:

  • text: If you help the needy, God will reward you.
hypothesis: Giving money to a poor man has no consequences.

An example of a non-TE (text does not entail nor contradict) is:

  • text: If you help the needy, God will reward you.
hypothesis: Giving money to a poor man will make you a better person.

Ambiguity of natural language

edit

A characteristic of natural language is that there are many different ways to state what one wants to say: several meanings can be contained in a single text and the same meaning can be expressed by different texts. This variability of semantic expression can be seen as the dual problem of language ambiguity. Together, they result in a many-to-many mapping between language expressions and meanings. The task of paraphrasing involves recognizing when two texts have the same meaning and creating a similar or shorter text that conveys almost the same information. Textual entailment is similar[6] but weakens the relationship to be unidirectional. Mathematical solutions to establish textual entailment can be based on the directional property of this relation, by making a comparison between some directional similarities of the texts involved.[4]

Approaches

edit

Textual entailment measures natural language understanding as it asks for a semantic interpretation of the text, and due to its generality remains an active area of research. Many approaches and refinements of approaches have been considered, such as word embedding, logical models, graphical models, rule systems, contextual focusing, and machine learning.[6] Practical or large-scale solutions avoid these complex methods and instead use only surface syntax or lexical relationships, but are correspondingly less accurate.[3] As of 2005, state-of-the-art systems are far from human performance; a study found humans to agree on the dataset 95.25% of the time.[7] Algorithms from 2016 had not yet achieved 90%.[8]

Applications

edit

Many natural language processing applications, like question answering, information extraction, summarization, multi-document summarization, and evaluation of machine translation systems, need to recognize that a particular target meaning can be inferred from different text variants. Typically entailment is used as part of a larger system, for example in a prediction system to filter out trivial or obvious predictions.[9] Textual entailment also has applications in adversarial stylometry, which has the objective of removing textual style without changing the overall meaning of communication.[10]

Datasets

edit

Some of available English NLI datasets include:

In addition, there are several non-English NLI datasets, as follows:

See also

edit

References

edit
  1. ^ Ido Dagan, Oren Glickman and Bernardo Magnini. The PASCAL Recognising Textual Entailment Challenge, p. 2 Archived 2012-03-03 at the Wayback Machine in: Quiñonero-Candela, J.; Dagan, I.; Magnini, B.; d'Alché-Buc, F. (Eds.) Machine Learning Challenges. Lecture Notes in Computer Science, Vol. 3944, pp. 177–190, Springer, 2006.
  2. ^ Korman, Daniel Z.; Mack, Eric; Jett, Jacob; Renear, Allen H. (2018-03-09). "Defining textual entailment". Journal of the Association for Information Science and Technology. 69 (6): 763–772. doi:10.1002/asi.24007. ISSN 2330-1635. S2CID 46920779.
  3. ^ a b Dagan, I. and O. Glickman. 'Probabilistic textual entailment: Generic applied modeling of language variability' Archived 2012-03-29 at the Wayback Machine in: PASCAL Workshop on Learning Methods for Text Understanding and Mining (2004) Grenoble.
  4. ^ a b Tătar, D. e.a. Textual Entailment as a Directional Relation
  5. ^ Textual Entailment Portal on the Association for Computational Linguistics wiki
  6. ^ a b Androutsopoulos, Ion; Malakasiotis, Prodromos (2010). "A Survey of Paraphrasing and Textual Entailment Methods" (PDF). Journal of Artificial Intelligence Research. 38: 135–187. arXiv:0912.3747. doi:10.1613/jair.2985. S2CID 9234833. Archived from the original (PDF) on 9 December 2017. Retrieved 13 February 2017.
  7. ^ Bos, Johan; Markert, Katja (6–8 October 2005). "Recognising textual entailment with logical inference". In Raymond Mooney; Joyce Chai; et al. (eds.). Proceedings of the conference on Human Language Technology and Empirical Methods in Natural Language Processing – HLT '05. Vancouver: Association for Computational Linguistics. pp. 628–635. doi:10.3115/1220575.1220654. S2CID 10202504.
  8. ^ Zhao, Kai; Huang, Liang; Ma, Mingbo (4 January 2017). "Textual Entailment with Structured Attentions and Composition". arXiv:1701.01126 [cs.CL].
  9. ^ Shani, Ayelett (25 October 2013). "How Dr. Kira Radinsky Used Algorithms to Predict Riots in Egypt". Haaretz. Retrieved 13 February 2017.
  10. ^ Potthast, Hagen & Stein 2016, p. 11-12.
  11. ^ Bowman, Samuel R.; Angeli, Gabor; Potts, Christopher; Manning, Christopher D. (2015). A large annotated corpus for learning natural language inference (PDF). In Proceedings of the 2015 Conference on Empirical Methods in Natural Language Processing (EMNLP). Association for Computational Linguistics. pp. 632–642. doi:10.18653/v1/D15-1075.
  12. ^ Williams, Adina; Nangia, Nikita; Bowman, Samuel R. (2018). A Broad-Coverage Challenge Corpus for Sentence Understanding through Inference (PDF). In Proceedings of the 2018 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 1 (Long Papers). Association for Computational Linguistics. pp. 1112–1122. doi:10.18653/v1/N18-1101.
  13. ^ Khot, Tushar; Sabharwal, Ashish; Clark, Peter (2018). "SciTaiL: A Textual Entailment Dataset from Science Question Answering". Proceedings of the AAAI Conference on Artificial Intelligence. 32 (1). doi:10.1609/aaai.v32i1.12022.
  14. ^ Marelli, Marco; Bentivogli, Luisa; Baroni, Marco; Bernardi, Raffaella; Menini, Stefano; Zamparelli, Roberto (2014). SemEval-2014 Task 1: Evaluation of Compositional Distributional Semantic Models on Full Sentences through Semantic Relatedness and Textual Entailment (PDF). In Proceedings of the 8th International Workshop on Semantic Evaluation (SemEval 2014). Dublin, Ireland: Association for Computational Linguistics. pp. 1–8. doi:10.3115/v1/S14-2001.
  15. ^ Romanov, Alexey; Shivade, Chaitanya (2018). Lessons from Natural Language Inference in the Clinical Domain (PDF). In Proceedings of the 2018 Conference on Empirical Methods in Natural Language Processing. Brussels, Belgium: Association for Computational Linguistics. pp. 1586–1596. doi:10.18653/v1/D18-1187.
  16. ^ Demszky, Dorottya; Guu, Kelvin; Liang, Percy (2018). "Transforming Question Answering Datasets Into Natural Language Inference Datasets". arXiv:1809.02922 [cs.CL].
  17. ^ Conneau, Alexis; Rinott, Ruty; Lample, Guillaume; Williams, Adina; Bowman, Samuel R.; Schwenk, Holger; Stoyanov, Veselin (2018). XNLI: Evaluating Cross-lingual Sentence Representations (PDF). In Proceedings of the 2018 Conference on Empirical Methods in Natural Language Processing. Brussels, Belgium: Association for Computational Linguistics. pp. 2475–2485. doi:10.18653/v1/D18-1269.
  18. ^ Amirkhani, Hossein; AzariJafari, Mohammad; Faridan-Jahromi, Soroush; Kouhkan, Zeinab; Pourjafari, Zohreh; Amirak, Azadeh (2023-07-07). "FarsTail: a Persian natural language inference dataset". Soft Computing. arXiv:2009.08820. doi:10.1007/s00500-023-08959-3. ISSN 1433-7479. S2CID 221802461.
  19. ^ Hu, Hai; Richardson, Kyle; Xu, Liang; Li, Lu; Kübler, Sandra; Moss, Lawrence (2020). OCNLI: Original Chinese Natural Language Inference (PDF). In Findings of the Association for Computational Linguistics: EMNLP 2020. pp. 3512–3526. doi:10.18653/v1/2020.findings-emnlp.314.
  20. ^ Wijnholds, Gijs; Moortgat, Michael (2021). SICK-NL: A Dataset for Dutch Natural Language Inference (PDF). In Proceedings of the 16th Conference of the European Chapter of the Association for Computational Linguistics: Main Volume. Association for Computational Linguistics. pp. 1474–1479. doi:10.18653/v1/2021.eacl-main.126.
  21. ^ Mahendra, Rahmad; Aji, Alham Fikri; Louvan, Samuel; Rahman, Fahrurrozi; Vania, Clara (2021). IndoNLI: A Natural Language Inference Dataset for Indonesian (PDF). In Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing. Association for Computational Linguistics. pp. 10511–10527. doi:10.18653/v1/2021.emnlp-main.821.

Bibliography

edit
edit