Researchers have introduced FINDERS, a novel framework that leverages diffusion language models (dLLMs) to create more adaptive user representations. Unlike existing methods that compress user behavior into static formats, FINDERS utilizes the iterative denoising process of dLLMs to capture multi-granularity information. The framework includes a router that dynamically emphasizes relevant behavioral factors based on specific scenario queries, enabling more effective cross-source fusion and compression. Experiments across ten industrial scenarios, including the construction of a new BehaviorQA corpus, show FINDERS outperforms current state-of-the-art methods in both unified and scenario-adaptive settings. AI
IMPACT This research could lead to more personalized and effective user modeling in various applications by improving how behavioral data is processed and utilized.
RANK_REASON The cluster contains a research paper detailing a new framework and methodology. [lever_c_demoted from research: ic=1 ai=1.0]
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →
