The ISO Standard for Dialogue Act Annotation, Second Edition

What it is about

The paper describes what changed in the second edition of ISO 24617-2, the multidimensional dialogue act scheme. Dependence relations become more precise: an answer or agreement has a functional dependence on an earlier act, while feedback has a feedback dependence on a "reference segment" that may be smaller than a whole utterance (one misheard word). A Contact Management dimension is added. The main novelty is the "layered plug-in": a plug-in brings its own abstract syntax, concrete syntax and semantics, so a dialogue act can be linked to its semantic content (attribute-value pairs, or events with semantic roles after ISO 24617-4), to a rhetorical relation with argument roles and a semantic/pragmatic domain (19 relations after DR-Core), to extra social-talk functions (politeness question, opening/closing politeness statement, farewell wish), to five-level feedback functions, or to an emotion with holder and object. No evaluation numbers; it is a specification paper.

Ideas

• I1. Give a responsive Say frame a functional dependence: an answer, agreement, disagreement, acceptance or refusal points to the frame of the question, offer or request it responds to (a dep index next to cause/next in prag/src/frames.rs). Then «What did the fox answer?» or «Did the crow agree?» can follow the link. Measure: speaker/says F1 on the frame bench and the count of questions about replies right in world knowledge diagnose.
• I2. Report act accuracy at two grains, as the ISO hierarchy allows: the 16 Act values and their ISO parents (Inform: assert/narrate/evaluate/express; Question: ask; Directive: request/suggest/warn; Commissive: offer/promise; Responsive: agree/disagree; SOM: thank/apologize/greet). Most of the perceptron's act errors (evaluate→assert 139, express→assert 50) stay inside Inform. Add a parent table to prag/src/schema.rs and a coarse column to prag eval; measure coarse vs fine macro-F1 on the silver.
• I3. Split the cause link by domain, as the second edition does for rhetorical relations: a semantic cause («Ray is sick; a virus attacked him») vs a pragmatic/epistemic one («Ray is sick; he beats his wife» — a reason for saying it). A cue: a cause clause attached to a Say frame, to an evaluation or to a modal of inference (must) is pragmatic. Measure cause precision on the frame bench before and after, and whether Ask::Cause answers improve.
• I4. Attach an emotion to express/evaluate acts as a triple (holder = the speaker, object = the frame evaluated, category/polarity). This makes polarity say whose evaluation of what, which the sentence-level label cannot. Measure polarity macro-F1 in prag eval when the object is known vs unknown.

For Mova

31. Speech acts §7.2–7.3 (already maps Act to ISO; the hierarchy and dependence are not used yet) and 38-dialogue-and-discourse-markers.md §10; 36-discourse-relations.md for the semantic/pragmatic cause. Code: prag/src/schema.rs (parent table), prag/src/frames.rs (functional dependence, cause domain), prag/src/eval.rs (coarse act score), world/src/talk.rs (the social-talk functions are already there). Measures: frame bench, prag eval act and polarity macro-F1.

Topics: Speech act theory

Cite

The paper itself, not this card. BibTeX from the library's record of the work.

@inproceedings{bunt2020iso,
  title = {{The ISO Standard for Dialogue Act Annotation, Second Edition}},
  author = {Bunt, Harry and Petukhova, Volha and Gilmartin, Emer and Pelachaud, Catherine and Fang, Alex and Keizer, Simon and Prévot, Laurent},
  booktitle = {Proceedings of the Twelfth Language Resources and Evaluation Conference},
  year = {2020},
  url = {https://aclanthology.org/2020.lrec-1.69/}
}

References

32 references, 2 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

1. Baggia, P., C. Pelachaud, C. Peter, and E. Zovato (2014). Emotion Markup Language (EmotionML) 1.0, W3C Recommendation 22 May 2014. http://www.w3.org/TR/2014/REC-emotionml-20140522/. ↗
2. Bunt, H. (2009a). The DIT++ taxonomy for functional dialogue markup. In D. Heylen, C. Pelachaud, R. Catizone, and D. Traum (Eds.), Proceedings of AAMAS-EDAML Workshop “Towards a Standard Markup Language for Embodied Dialogue Acts, Budapest”, pp. 36–36. find
3. Bunt, H. (2012). The semantics of feedback. In Proceedings of SeineDial, 16th Workshop on the Semantics and Pragmatics of Dialogue, Paris, pp. 118–127. find
4. Bunt, H. (2013). A context-change semantics for dialogue acts. In H. Bunt, J. Bos, and S. Pulman (Eds.), Computing Meaning, vol. 4, pp. 177–201. Dordrecht: Springer. find
5. Bunt, H. (2019). Plug-ins for content annotation of dialogue acts. In Proceedings 15th Joint ISO-ACL Workshop on Interoperable Semantic Annotation (ISA-15), Gothenburg, Sweden, pp. 34–45. find
6. Bunt, H. (2020). The annotation of quantification: The current state of ISO 24617-12. In Proceedings of the 16th Joint ISO-ACL Workshop on Interoperable Semantic Annotation (ISA-16), Marseille. find
7. Bunt, H., V. Petukhova, A. Malchanau, A. Fang, and K. Wijnhoven (2019). The DialogBank: Dialogues with interoperable annotations. Language Resources and Evaluation 53, 213–249. Available online at DOI: 10.1007/s10579-018-9436-9. ↗
8. Bunt, H., J. Pustejovsky, and K. Lee (2018). Towards an ISO Standard for the Annotation of Quantification. In Proceedings of the 11th International Conference on Language Resources and Evaluation (LREC 2018), Miyazaki, Japan. find
9. Bunt, H. and L. Romary (2004). Standardization in Multimodal Content Representation: Some methodological issues. In Proceedings of 4th International Conference on Language Resources and Evaluation (LREC 2004), Lisbon, pp. 2219–2222. Paris: ELRA. find
10. Burkhardt, F., C. Pelachaud, B. Schuller, and E. Zovato (2017). EmotionML. In D. Dahl (Ed.), Multimodal Interaction with W3C Standards, pp. 65–80. Cham (Switzerland): Springer. find
11. Chowdhury, S., E. Stepanov, and G. Riccardi (2014). Transfer of corpus-specific dialog act annotation to the iso standard: Is it worth it? In Proceedings of the 9th International Conference on Language Resources and Evaluation (LREC 2014), Paris. ELRA. find
12. Crible, L. and L. Degand (2019). Domains and functions: A two-dimensional account of discourse markers. Discours 24. Available online at DOI: 10.400/dis-cours.9997. find
13. Ekman, P. (1972). Universals and cultural differences in facial expressions of emotion. In J. Cole (Ed.), Nebraska Symposium on Motivation, Vol. 19, pp. 207–282. University of Nebraska Press. find
14. Fang, A., J. Cao, H. Bunt, and X. Liu (2012). The annotation of the Switchboard corpus with the new ISO standard for dialogue act analysis. In Proceedings 8th Joint ISO-ACL SIGSEM Workshop on Interoperable Semantic Annotation (ISA-8), ILC-CNR, Pisa. find
15. Geertzen, J., V. Petukhova, and H. Bunt (2008). Evaluating dialogue act tagging with naive and expert annotators. In Proceedings of the 6th Inernational Conference on Language Resources and Evaluation (LREC 2008), Marrakech. Paris: ELRA. find
16. Gilmartin, E., C. Saam, B. Spillane, M. O’Reilly, K. Su, A. Calvo, L. Cerrato, K. Levacher, N. Campbell, and V. Wade (2018). The ADELE corpus of dyadic social text conversations: Dialogue act annotation with ISO 24617-2. In Proceedings 11th International Conference on Language Resources and Evaluation (LREC 2018). find
17. Ide, N. and H. Bunt (2010). Anatomy of annotation schemes: Mapping to GrAF. In Proceedings 4th Linguistic Annotation Workshop (LAW IV), Uppsala., pp. 247–255. find
18. Ide, N. and L. Romary (2004). International Standard for a Linguistic Annotation Framework. Natural Language Engineering 10, 211–225. find
19. ISO (2012a). ISO 24617-1: 2012, Language Resource Management - Semantic Annotation Framework (SemAF) - Part 1: Time and events. Geneva: International Organisation for Standardisation ISO. find
20. ISO (2012b). ISO 24617-2:2012, Language Resource Management - Semantic Annotation Framework (SemAF) - Part 2: Dialogue acts. Geneva: International Organisation for Standardisation ISO. find
21. ISO (2014). ISO 24617-4: 2014, Language Resource Management - Semantic Annotation Framework (SemAF) - Part 4: Semantic roles. Geneva: International Organisation for Standardisation ISO. find
22. ISO (2015). ISO 24617-6:2015, Language Resource Management - Semantic Annotation Framework (SemAF) - Part 6: Principles of semantic annotation. Geneva: International Organisation for Standardisation ISO. find
23. ISO (2016). ISO 24617-8:2016, Language Resource Management - Semantic Annotation Framework (SemAF) - Part 8: Semantic relations in discourse, Core annotation scheme (DR-Core). Geneva: International Organisation for Standardisation ISO. find
24. Keizer, S. and H. Bunt (2006). Multidimensional dialogue management. In Proceedings 7th SIGdial Workshop on Discourse and Dialogue, Sydney, Australia, pp. 37–45. find
25. Keizer, S., H. Bunt, and V. Petukhova (2011). Multidimensional dialogue management. In A. van den Bosch and G. Bouma (Eds.), Interactive Multimodal Question Answering, pp. 57–86. Berlin: Springer. find
26. Keizer, S., O. Dusek, X. Liu, and V. Rieser (2019). User evaluation of a multi-dimensional statistical dialogue system. In Proceedings 20th SIG. find
27. Malchanau, A. (2019). Cognitive Architecture for Multimodal Multidimensional Dialogue Management. Saarbrücken: PhD Thesis, University of Saarland. find
28. Malchanau, A., V. Petukhova, and H. Bunt (2019). Towards integration of cognitive models in dialogue management: Designing the virtual negotiation coach application. Dialogue and Discourse 9(2), 35–79. find
29. Mezza, S., A. Cervone, E. Stepanov, G. Tortoretto, and G. Riccardi (2018). ISO-standard domain-independent dialogue act tagging for conversational agents. In Proceedings of he 27th International Conference on Computational Linguistics (COLING 2018), Santa Fé, New Mexico, USA, pp. 3539–3551. Association for Computational Linguistics. find our card
30. Ngo, T., K. Pham, and H. Takeda (2018). A Vietnamese dialogue act corpus based on the ISO 24617-2 standard. In Proceedings of the 11th International Conference on Language Resources and Evaluation (LREC 2018), Miyazaki, Japan. Paris: ELRA. find
31. Petukhova, V. and H. Bunt (2009). Grounding by nodding. In Proceedings GESPIN, Conference on Gestures and Speech in Interaction, Poznán. find
32. Prasad, R., N. Dinesh, A. Lee, E. Miltsakaki, L. Robaldo, A. Joshi, and B. Webber (2008). The Penn Discourse TreeBank 2.0. In Proceedings 6th International Conference on Language Resources and Evaluation (LREC 2008), Marrakech. Paris: ELRA. find our card

This page in Markdown → https://md.mova.elamur.ai/free/en/docs/papers/bunt-2020-iso-standard-dialogue-act-annotation/