Learning to Retrieve User History and Generate User Profiles for Personalized Persuasiveness Prediction
Fuente:
arXiv
Saved in:
| Main Authors: | , , , |
|---|---|
| Format: | Preprint |
| Published: |
2026
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866917419468980224 |
|---|---|
| author | Park, Sejun Park, Yoonah Lim, Jongwon Jo, Yohan |
| author_facet | Park, Sejun Park, Yoonah Lim, Jongwon Jo, Yohan |
| contents | Estimating the persuasiveness of messages is critical in various applications, from recommender systems to safety assessment of LLMs. While it is imperative to consider the target persuadee's characteristics, such as their values, experiences, and reasoning styles, there is currently no established systematic framework to optimize leveraging a persuadee's past activities (e.g., conversations) to the benefit of a persuasiveness prediction model. To address this problem, we propose a context-aware user profiling framework with two trainable components: a query generator that generates optimal queries to retrieve persuasion-relevant records from a user's history, and a profiler that summarizes these records into a profile to effectively inform the persuasiveness prediction model. Our evaluation on the ChangeMyView Reddit dataset shows consistent improvements over existing methods across multiple predictor models, raising F1 from 33% to 47% on Llama-3.3-70B-Instruct. Further analysis shows that effective user profiles are context-dependent and predictor-specific, rather than relying on static attributes or surface-level similarity. Together, these results highlight the importance of task-oriented, context-dependent user profiling for personalized persuasiveness prediction. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2601_05654 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Learning to Retrieve User History and Generate User Profiles for Personalized Persuasiveness Prediction Park, Sejun Park, Yoonah Lim, Jongwon Jo, Yohan Computation and Language Artificial Intelligence Estimating the persuasiveness of messages is critical in various applications, from recommender systems to safety assessment of LLMs. While it is imperative to consider the target persuadee's characteristics, such as their values, experiences, and reasoning styles, there is currently no established systematic framework to optimize leveraging a persuadee's past activities (e.g., conversations) to the benefit of a persuasiveness prediction model. To address this problem, we propose a context-aware user profiling framework with two trainable components: a query generator that generates optimal queries to retrieve persuasion-relevant records from a user's history, and a profiler that summarizes these records into a profile to effectively inform the persuasiveness prediction model. Our evaluation on the ChangeMyView Reddit dataset shows consistent improvements over existing methods across multiple predictor models, raising F1 from 33% to 47% on Llama-3.3-70B-Instruct. Further analysis shows that effective user profiles are context-dependent and predictor-specific, rather than relying on static attributes or surface-level similarity. Together, these results highlight the importance of task-oriented, context-dependent user profiling for personalized persuasiveness prediction. |
| title | Learning to Retrieve User History and Generate User Profiles for Personalized Persuasiveness Prediction |
| topic | Computation and Language Artificial Intelligence |
| url | https://arxiv.org/abs/2601.05654 |