Individual and Domain Adaptation in Sentence Planning for Dialogue
By Marilyn A. Walker, Amanda Stent, François Mairesse, Rashmi Prasad. 2007
165 citations recorded by the source corpus.
One of the biggest challenges in the development and deployment of spoken dialogue systems is the design of the spoken language generation module. This challenge arises from the need for the generator to adapt to many features of the dialogue domain, user population, and dialogue context. A promising approach is trainable generation, which uses general-purpose linguistic knowledge that is automatically adapted to the features of interest, such as the application domain, individual user, or user group. In this paper we present and evaluate a trainable sentence planner for providing restaurant information in the MATCH dialogue system. We show that trainable sentence planning can produce complex information presentations whose quality is comparable to the output of a template-based generator tuned to this domain. We also show that our method easily supports adapting the sentence planner to individuals, and that the individualized sentence planners generally perform better than models trained and tested on a population of individuals. Previous work has documented and utilized individual preferences for content selection, but to our knowledge, these results provide the first demonstration of individual preferences for sentence planning operations, affecting the content order, discourse structure and sentence structure of system responses. Finally, we evaluate the contribution of different feature sets, and show that, in our application, n-gram features often do as well as features based on higher-level linguistic representations.
Subfields: Artificial Intelligence, Dialogue & Conversational AI
How researchers see this paper
The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.
In Artificial Intelligence
- Ondrej Dusek
- Verena Rieser
- Jianfeng Gao
- Jekaterina Novikova
- Mohit Bansal
- Percy Liang
- Richard Socher
- Caiming Xiong
In Dialogue & Conversational AI
- Verena Rieser
- Pei-hao Su
- Susannah Young
- Milica Gasic
- Tsung-Hsien Wen
- Oliver Lemon
- Nikola Mrksic
- Emiel Krahmer
Citations in the corpus
- Neural Approaches to Conversational AI (2018, neutral): “Even though machine learning can be used to train certain parts of these systems (Langkilde and Knight, 1998; Stent et al., 2004; Walker et al., 2007), the cost to write and maintain templates and rules leads to challenges in adapting to new domains or different user populations.”
- Semantically Conditioned LSTM-based Natural Language Generation for Spoken Dialogue Systems (2015, background): “These systems include specific trainable modules within the generation framework to allow the model to adapt to different domains (Walker et al., 2007), or reproduce certain style (Mairesse and Walker, 2011).”
- Survey of the State of the Art in Natural Language Generation: Core tasks, applications and evaluation (2017, background): “For example, Stent and Molina (2009) describe an approach to the acquisition of sentence-combining rules from a discourse treebank, which are then incorporated into the sparky sentence planner described by Walker et al. (2007b).”
- Controlling User Perceptions of Linguistic Style: Trainable Generation of Personality Traits (2011, methodology): “Statistical natural language generation (SNLG) has the potential to address such scalability issues by relying on annotated data rather than manual parameter tuning.”
- RankME: Reliable Human Ratings for Natural Language Generation (2018, background): “001 ) for the same utterance (Walker et al. , 2007).”
- Relevance of Unsupervised Metrics in Task-Oriented Dialogue for Evaluating Natural Language Generation (2017, methodology): “Other statistical approaches were proposed us-ing tree-based models and reinforcement learning (Walker et al., 2007; Rieser and Lemon, 2009).”
- Natural Language Generation as Planning Under Uncertainty for Spoken Dialogue Systems (2009, methodology): “sentence plans can be adapted to the individual user and the domain [ 29 ].”
- Constrained Decoding for Neural NLG from Compositional Representations in Task-Oriented Dialogue (2019, methodology): “While several studies have shown that controlling such discourse behaviors can be critical to user perceptions of quality and naturalness (Lemon et al., 2004; Carenini and Moore, 2006; Walker et al., 2007; White et al., 2010; Demberg et al., 2011), flat MRs provide no means to do so.”
- End-to-End Content and Plan Selection for Data-to-Text Generation (2018, background): “These representations often model the hierarchical structure of discourse relations (Walker et al., 2007).”
- Natural Language Generation as Incremental Planning Under Uncertainty: Adaptive Information Presentation for Statistical Dialogue Systems (2014, methodology): “[30], [31], while much work in SDS uses a RECOMMEND strategy, e.”
- The Extended SPaRKy Restaurant Corpus: Designing a Corpus with Variable Information Density (2017, methodology): “the MATCH corpus, [9]) includes more complex meaning representations based on Rhetorical Structure Theory (RST, [16]).”
- Computational Generation of Referring Expressions: A Survey (2012, methodology): “concentrating on choice optimization during planning based on user data (Walker et al. 2007; White, Clark, and Moore 2010).”
- Can Neural Generators for Dialogue Learn Sentence Planning and Discourse Structuring? (2018, background): “…selection and linearization (Marcu, 1997; Lapata, 2003; Barzilay and Lap-ata, 2005), and trainable sentence planning for discourse structure and aggregation operations (Stent and Molina, 2009; Walker et al., 2007; Paiva and Evans, 2004; Sauper and Barzilay, 2009; H. Cheng and Mellish, 2001).”
- Policy-Driven Neural Response Generation for Knowledge-Grounded Dialog Systems (2020, background): “There has been previous work on task-oriented systems that proposed explicit content and sentence planning (Walker et al., 2007) to further control the”