Towards Stable and Personalised Profiles for Lexical Alignment in Spoken Human-Agent Dialogue
Keara Schaaij, Roel Boumans, Tibor Bosse, Iris Hendrickx
TL;DR
The study addresses the challenge of enabling lexical alignment in human-agent dialogue by constructing stable, personalised lexical profiles from spontaneous Dutch interviews. It systematically varies data size and POS-item counts, and evaluates temporal stability using recall, coverage, and cosine similarity across held-out data, finding that profiles built from the first ~10 minutes with a compact POS-item set yield the best data efficiency and robustness. The results demonstrate stable reuse of profile content over time and reasonable generalisation to unseen data, despite low overall coverage reflecting the limited profile scope. This work provides a practical foundation for lexical alignment in agents when real-time user input is limited, with implications for inclusive interactions such as dementia care and directions for future validation and adaptive updating of profiles.
Abstract
Lexical alignment, where speakers start to use similar words across conversation, is known to contribute to successful communication. However, its implementation in conversational agents remains underexplored, particularly considering the recent advancements in large language models (LLMs). As a first step towards enabling lexical alignment in human-agent dialogue, this study draws on strategies for personalising conversational agents and investigates the construction of stable, personalised lexical profiles as a basis for lexical alignment. Specifically, we varied the amounts of transcribed spoken data used for construction as well as the number of items included in the profiles per part-of-speech (POS) category and evaluated profile performance across time using recall, coverage, and cosine similarity metrics. It was shown that smaller and more compact profiles, created after 10 min of transcribed speech containing 5 items for adjectives, 5 items for conjunctions, and 10 items for adverbs, nouns, pronouns, and verbs each, offered the best balance in both performance and data efficiency. In conclusion, this study offers practical insights into constructing stable, personalised lexical profiles, taking into account minimal data requirements, serving as a foundational step toward lexical alignment strategies in conversational agents.