Textual entailment

From Wikipedia, the free encyclopedia
Jump to: navigation, search

Textual entailment (TE) in natural language processing is a directional relation between text fragments. The relation holds whenever the truth of one text fragment follows from another text. 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] The relation is directional because even if "t entails h", the reverse "h entails t" is much less certain.[2][3]

Ambiguity of natural language[edit]

A characteristic of natural language is that there are many different ways to state what you want to say: several meanings can be contained in a single text and that 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. Interpreting a text correctly would, in theory, require a thorough semantic interpretation into a logic-based representation of its meanings. Practical solutions for natural language processing seek to go not that deep and use textual entailment in a more shallow way.[2]

Examples[edit]

Textual entailment can be illustrated with examples of three different relations:[4]

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 better person.

Recognizing textual entailment[edit]

Many natural language processing applications, like Question Answering (QA), Information Extraction (IE), (multi-document) summarization and machine translation (MT) evaluation, need a model for this variability phenomenon in order to recognize that a particular target meaning can be inferred from different text variants. In 2004 Recognizing Textual Entailment (RTE) has been proposed as a generic task that captures major semantic inference needs across many natural language processing applications.[2] From 2004 to 2013 eight RTE Challenges were organized with the aim of providing researchers with concrete datasets on which to evaluate and compare their approaches. The main organizers of the RTE Challenges along the years were Bar-Ilan University, Fondazione Bruno Kessler, CELCT, and NIST.

Mathematical solutions to establish textual entailment can be based on the directional property of this relation, by making a comparison between some directional similarity of the texts involved.[3]

Notes[edit]

  1. ^ Ido Dagan, Oren Glickman and Bernardo Magnini. The PASCAL Recognising Textual Entailment Challenge, p. 2 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. ^ a b c Dagan, I. and O. Glickman. 'Probabilistic textual entailment: Generic applied modeling of language variability' in: PASCAL Workshop on Learning Methods for Text Understanding and Mining (2004) Grenoble.
  3. ^ a b Tătar, D. e.a. Textual Entailment as a Directional Relation
  4. ^ Textual Entailment Portal on the [Association for Computational Linguistics] wiki