Rhetorical Agreement: Maintaining Cohesive Conversations

  • Boris Galitsky


To support a natural flow of a conversation in a chatbot, rhetorical structures of each message has to be analyzed. We classify a pair of paragraphs of text as appropriate for one to follow another, or inappropriate, based on communicative discourse considerations. To represent a multi-sentence message with respect to how it should follow a previous message in a conversation or dialogue, we build an extension of a discourse tree for it. Extended discourse tree is based on a discourse tree for RST relations with labels for communicative actions, and also additional arcs for anaphora and ontology-based relations for entities. We refer to such trees as Communicative Discourse Trees (CDTs). We explore syntactic and discourse features that are indicative of correct vs incorrect request-response or question-answer pairs. Two learning frameworks are used to recognize such correct pairs: deterministic, nearest-neighbor learning of CDTs as graphs, and a tree kernel learning of CDTs, where a feature space of all CDT sub-trees is subject to SVM learning. We form the positive training set from the correct pairs obtained from Yahoo Answers, social network, corporate conversations including Enron emails, customer complaints and interviews by journalists. The corresponding negative training set is artificially created by attaching responses for different, inappropriate requests that include relevant keywords. The evaluation showed that it is possible to recognize valid pairs in 70% of cases in the domains of weak request-response agreement and 80% of cases in the domains of strong agreement, which is essential to support automated conversations. These accuracies are comparable with the benchmark task of classification of discourse trees themselves as valid or invalid, and also with classification of multi-sentence answers in factoid question-answering systems. The applicability of proposed machinery to the problem of chatbots, social chats and programming via NL is demonstrated. We conclude that learning rhetorical structures in the form of CDTs is the key source of data to support answering complex questions, chatbots and dialogue management.


  1. Airenti G, Bara BG, Colombetti M (1993) Conversation and behavior games in the pragmatics of dialogue. Cogn Sci 17:197–256CrossRefGoogle Scholar
  2. Allen J, Perrault C (1980) Analyzing intention in utterances. Artif Intell 15(3):143–178CrossRefGoogle Scholar
  3. Baumeister RF, Bushman BJ (2010) Social psychology and human nature: international edition. Wadsworth, BelmontGoogle Scholar
  4. Bengio Y, Ducharme R, Vincent P, Janvin C (2003) A neural probabilistic language model. J Mach Learn Res 3(March 2003):1137–1155zbMATHGoogle Scholar
  5. Blaylock N, Allen J, Ferguson G (2003) Managing communicative intentions with collaborative problem solving. In: Current and new directions in discourse and dialogue. Springer Netherlands, Dordrecht, pp 63–84CrossRefGoogle Scholar
  6. Burstein JC, Braden-Harder L, Chodorow MS, Kaplan BA, Kukich K, Lu C, Rock DA, Wolff S (2002) System and method for computer-based automatic essay scoring. United States Patent 6,366,759: Educational Testing ServiceGoogle Scholar
  7. Cohen W (2016) Enron email dataset. Last downloaded 10 July 2016
  8. Cohen PR, Levesque HJ (1990) Intention is choice with commitment. Artif Intell 42:213–261MathSciNetCrossRefGoogle Scholar
  9. Collins M, Duffy N (2002) Convolution kernels for natural language. In: Proceedings of NIPS, pp 625–632Google Scholar
  10. Coulthard RM, Brazil D (1979) Exchange structure: discourse analysis monographs no. 5. The University of Birmingham, English Language Research, BirminghamGoogle Scholar
  11. Cristea D, Ide N, Romary L (1998) Veins theory: a model of global discourse cohesion and coherence. In: Boitet C, Whitelock P (eds) 17th international conference on computational linguistics, vol 1. Association for Computational Linguistics, Montreal, pp 281–285Google Scholar
  12. De Boni M (2007) Using logical relevance for question answering. J Appl Log 5(1):92–103MathSciNetCrossRefGoogle Scholar
  13. Dijkstra EW (1965) Programming considered as a human activity. In: Proceedings of the IFIP Congress, pp 213–217Google Scholar
  14. Galitsky B (2013) Machine learning of syntactic parse trees for search and classification of text. Eng Appl Artif Intell 26(3):1072–1091CrossRefGoogle Scholar
  15. Galitsky B (2016) Using extended tree kernels to recognize metalanguage in text. In: Kreinovich V (ed) Uncertainty modeling. Springer, ChamGoogle Scholar
  16. Galitsky B (2017) Matching parse thickets for open domain question answering. Data Knowl Eng 107:24–50CrossRefGoogle Scholar
  17. Galitsky B, Ilvovsky D (2017) On a chatbot finding answers with optimal rhetoric representation. In: Proceedings of recent advances in natural language processing, pp 253–259Google Scholar
  18. Galitsky B, Lebedeva N (2015) Recognizing documents versus meta-documents by tree kernel learning. In: FLAIRS conference, pp 540–545Google Scholar
  19. Galitsky B, McKenna EW (2017) Sentiment extraction from consumer reviews for providing product recommendations. US Patent 9,646,078Google Scholar
  20. Galitsky B, Shpitsberg I (2016) Autistic learning and cognition. In: Computational autism. Springer, Cham, pp 245–293CrossRefGoogle Scholar
  21. Galitsky B, Usikov D (2008) Programming spatial algorithms in natural language. AAAI workshop technical report WS-08-11, Palo Alto, pp 16–24Google Scholar
  22. Galitsky B, Kuznetsov SO, Samokhin MV (2005) Analyzing conflicts with concept-based learning. In: International conference on conceptual structures, pp 307–322Google Scholar
  23. Galitsky B, Dobrocsi G, de la Rosa JL, Kuznetsov SO (2011) Using generalization of syntactic parse trees for taxonomy capture on the web. In: International conference on conceptual structures, pp 104–117Google Scholar
  24. Galitsky B, Usikov D, Kuznetsov SO (2013) Parse thicket representations for answering multi-sentence questions. In: 20th international conference on conceptual structures. ICCS, p 95Google Scholar
  25. Galitsky B, Ilvovsky D, Lebedeva N, Usikov D (2014) Improving trust in automation of social promotion. In: AAAI Spring symposium on the intersection of robust intelligence and trust in autonomous systems, Stanford CAGoogle Scholar
  26. Galitsky B, Ilvovsky D, Kuznetsov SO (2015a) Text integrity assessment: sentiment profile vs rhetoric structure. CICLing-2015, CairoGoogle Scholar
  27. Galitsky B, Ilvovsky D, Kuznetsov SO (2015b) Rhetoric map of an answer to compound queries. Knowledge Trail Inc. ACL 2015, Beijing, pp 681–686Google Scholar
  28. Galitsky B, Ilvovsky D, Kuznetsov SO (2015c) Text classification into abstract classes based on discourse structure. In: Proceedings of recent advances in natural language processing, Hissar, Bulgaria, 7–9 September 2015, pp 200–207Google Scholar
  29. Ganter B, Kuznetsov SO (2003) Hypotheses and Version Spaces, Proc. 10th Int. Conf. on Conceptual Structures, ICCS’03, Lecture Notes in Artificial Intelligence, vol 2746, pp 83–95Google Scholar
  30. Grefenstette E, Dinu G, Zhang Y, Sadrzadeh M and Baroni M (2013) Multi-step regression learning for compositional distributional semantics. In Proceedings of the Tenth International Conference on Computational Semantics. Association for Computational LinguisticsGoogle Scholar
  31. Grosz B, Sidner C (1986) Attention, intention, and the structure of discourse. Comput Linguist 12(3):175–204Google Scholar
  32. Jansen P, Surdeanu M, Clark P (2014) Discourse complements lexical semantics for nonfactoid answer reranking. In: Proceedings of the 52nd ACLGoogle Scholar
  33. Joty SR, Carenini G, Ng RT (2016) CODRA: a novel discriminative framework for rhetorical analysis. Comput Linguist 41(3):385–435MathSciNetCrossRefGoogle Scholar
  34. Jurafsky D, Martin JH (2000) Speech and language processing: an introduction to natural language processing, computational linguistics, and speech recognition. Prentice Hall, Upper Saddle RiverGoogle Scholar
  35. Kate R, Wong YW, Mooney R (2005) Learning to transform natural to formal languages. Proc Natl Conf Artif Intell 20:1062–1068Google Scholar
  36. Kipper K, Korhonen A, Ryant N, Palmer M (2008) A large-scale classification of English verbs. Language Resources and Evaluation Journal 42:21–40CrossRefGoogle Scholar
  37. Kontos J, Malagardi I, Peros J (2016) Question answering and rhetoric analysis of biomedical texts in the AROMA system. Unpublished manuscript. Last downloaded 12 September 2016
  38. Kuznetsov SO (1999) Learning of simple conceptual graphs from positive and negative examples. In: European conference on principles of data mining and knowledge discovery. Springer, Berlin/Heidelberg, pp 384–391CrossRefGoogle Scholar
  39. Levinson SC (2000) Presumptive meanings: the theory of generalized conversational implicature. The MIT Press, Cambridge, MACrossRefGoogle Scholar
  40. Litman DL, Allen JF (1987) A plan recognition model for subdialogues in conversation. Cogn Sci 11:163–200CrossRefGoogle Scholar
  41. Mann W, Thompson S (1988) Rhetorical structure theory: towards a functional theory of text organization. Text-Interdiscipl J Stud Discourse 8(3):243–281CrossRefGoogle Scholar
  42. Mikolov T, Chen K, Corrado GS, Jeffrey D (2015) Computing numeric representations of words in a high-dimensional space. US Patent 9,037,464, Google, Inc.Google Scholar
  43. Mitchell J, Lapata M (2010) Composition in distributional models of semantics. Cogn Sci 34(8):1388–1429CrossRefGoogle Scholar
  44. Mitocariu E, Anechitei DA, Cristea D (2016) Comparing discourse tree structures. Available from: Accessed 15 May 2016
  45. Moschitti A, Quarteroni S, Basili R, and Manandhar S (2007) Exploiting syntactic and shallow semantic kernels for question/answer classification. In ACL’07, Prague, Czech RepublicGoogle Scholar
  46. Peldszus A, Stede M (2013) From argument diagrams to argumentation mining in texts: a survey. Int J Cognit Informat Nat Intell 7(1):1–31CrossRefGoogle Scholar
  47. Popescu V, Caelen J, Burileanu C (2007) Logic-based rhetorical structuring for natural language generation in human-computer dialogue. Lect Notes Comput Sci 4629:309–317CrossRefGoogle Scholar
  48. Popescu-Belis A (2005) Dialogue acts: one or more dimensions? Tech report ISSCO working paper n. 62Google Scholar
  49. Radev DR, Jing H, Budzikowska M (2000) Centroid-based summarization of multiple documents: sentence extraction, utility-based evaluation, and user studies. In: Proceedings of the 2000 NAACL-ANLP workshop on automatic summarization, vol 4Google Scholar
  50. Reichman R (1985) Getting computers to talk like you and me: discourse context, focus and semantics (an ATN model). MIT Press, Cambridge, MA/LondonGoogle Scholar
  51. Santhosh S, Ali J (2012) Discourse based advancement on question answering system. J Soft Comput 1(2):1–12Google Scholar
  52. Schiffrin D (2005) Discourse. In: Dittmar N, Trudgill P (eds) Handbook of sociolinguistics. Mouton, de GruyterGoogle Scholar
  53. Scholman M, Evers-Vermeul J, Sanders T (2016) Categories of coherence relations in discourse annotation. Dialogue Discourse 7(2):1–28Google Scholar
  54. Socher RC, Manning D, Ng AY (2010) Learning continuous phrase representations and syntactic parsing with recursive neural networks. In: Proceedings of the NIPS-2010 deep learning and unsupervised feature learning workshopGoogle Scholar
  55. Sparck Jones K (1995) Summarising: analytic framework, key component, experimental method. In: Endres-Niggemeyer B, Hobbs J, Sparck Jones K (eds) Summarising text for intelligent communication, Dagstuhl seminar report 79, 13.12–17.12.93 (9350). Dagstuhl, WadernGoogle Scholar
  56. Sperber D, Wilson D (1986) Relevance: communication and cognition. Blackwell/Oxford/Harvard University Press, CambridgeGoogle Scholar
  57. Surdeanu M, Hicks T, Valenzuela-Escarcega MA (2015) Two practical rhetorical structure theory parsers. In: Proceedings of the conference of the North American chapter of the association for computational linguistics – human language technologies: software demonstrations (NAACL HLT)Google Scholar
  58. Traum DR, Hinkelman EA (1992) Conversation acts in task-oriented spoken dialogue. Comput Intell 8(3):575–599CrossRefGoogle Scholar
  59. Tsui AMB (1994) English conversation. Describing english language series. Oxford University Press, LondonGoogle Scholar
  60. Yessenalina A, Cardie C (2011) Compositional matrix-space models for sentiment analysis. In: EMNLP’11. Association for Computational Linguistics, Stroudsburg, pp 172–182Google Scholar
  61. Walker MA, Passonneau RJ, Boland JE (2001) Quantitative and qualitative evaluation of DARPA communicator spoken dialogue systems. In: Proceedings of the ACL, pp 515–522Google Scholar
  62. Wang W, Su J, Tan CL (2010) Kernel based discourse relation recognition with temporal ordering information. ACLGoogle Scholar
  63. Wilks YA (ed) (1999) Machine conversations. Kluwer, New YorkGoogle Scholar
  64. Zanzotto FM, Korkontzelos I, Fallucchi F, Manandhar S (2010) Estimating linear models for 2112 compositional distributional semantics. In Proceedings of the 23rd International Conference 2113 on Computational Linguistics (COLING)Google Scholar

Copyright information

© Springer Nature Switzerland AG 2019

Authors and Affiliations

  • Boris Galitsky
    • 1
  1. 1.Oracle (United States)San JoseUSA

Personalised recommendations