Consensus vs. Dissent: Dynamic LLM Modeling of Subjective Preferences in Group Recommenders

Name: Consensus vs. Dissent: Dynamic LLM Modeling of Subjective Preferences in Group Recommenders
Uploaded: 2026-10-10T07:34:17.398627+00:00
Duration: 474 s
Description: The research developed a system where Large Language Models (LLMs) dynamically choose the best recommendation strategy for group systems by mimicking human judgment. By fine-tuning LLMs on survey data, they created models that predict human-like fairness, satisfaction, and consensus scores. This dynamic approach outperforms static methods in user studies.

The research developed a system where Large Language Models (LLMs) dynamically choose the best recommendation strategy for group systems by mimicking human judgment. By fine-tuning LLMs on survey data, they created models that predict human-like fairness, satisfaction, and consensus scores. This dynamic approach outperforms static methods in user studies.

Open the full episode