GroundHog: Dialogue Generation using Multi-Grained Linguistic Input

Alexander Chernyavskiy, Lidiia Ostyakova, Dmitry Ilvovsky


Abstract
Recent language models have significantly boosted conversational AI by enabling fast and cost-effective response generation in dialogue systems. However, dialogue systems based on neural generative approaches often lack truthfulness, reliability, and the ability to analyze the dialogue flow needed for smooth and consistent conversations with users. To address these issues, we introduce GroundHog, a modified BART architecture, to capture long multi-grained inputs gathered from various factual and linguistic sources, such as Abstract Meaning Representation, discourse relations, sentiment, and grounding information. For experiments, we present an automatically collected dataset from Reddit that includes multi-party conversations devoted to movies and TV series. The evaluation encompasses both automatic evaluation metrics and human evaluation. The obtained results demonstrate that using several linguistic inputs has the potential to enhance dialogue consistency, meaningfulness, and overall generation quality, even for automatically annotated data. We also provide an analysis that highlights the importance of individual linguistic features in interpreting the observed enhancements.
Anthology ID:
2024.codi-1.14
Volume:
Proceedings of the 5th Workshop on Computational Approaches to Discourse (CODI 2024)
Month:
March
Year:
2024
Address:
St. Julians, Malta
Editors:
Michael Strube, Chloe Braud, Christian Hardmeier, Junyi Jessy Li, Sharid Loaiciga, Amir Zeldes, Chuyuan Li
Venues:
CODI | WS
SIG:
Publisher:
Association for Computational Linguistics
Note:
Pages:
149–160
Language:
URL:
https://aclanthology.org/2024.codi-1.14
DOI:
Bibkey:
Cite (ACL):
Alexander Chernyavskiy, Lidiia Ostyakova, and Dmitry Ilvovsky. 2024. GroundHog: Dialogue Generation using Multi-Grained Linguistic Input. In Proceedings of the 5th Workshop on Computational Approaches to Discourse (CODI 2024), pages 149–160, St. Julians, Malta. Association for Computational Linguistics.
Cite (Informal):
GroundHog: Dialogue Generation using Multi-Grained Linguistic Input (Chernyavskiy et al., CODI-WS 2024)
Copy Citation:
PDF:
https://aclanthology.org/2024.codi-1.14.pdf
Supplementary material:
 2024.codi-1.14.SupplementaryMaterial.zip
Video:
 https://aclanthology.org/2024.codi-1.14.mp4