Collaborative Editable Model

2025-06-17

AI Summary

Large language models (LLMs) for vertical domains face bottlenecks in requiring massive annotated data, high computational resources, and lengthy adaptation cycles. Method: This paper proposes a user-coconstructed zero-shot domain enhancement framework. It dynamically acquires user feedback via conversational interaction, employs attribution analysis and a scoring mechanism to identify high-value knowledge snippets, and injects them into context prompts in a structured manner—enabling fine-tuning-free, parameter-free dynamic domain evolution. A novel user collaborative editing mechanism is introduced to transform distributed, real-time feedback into interpretable and verifiable knowledge units. Results: Evaluated on 15k real-world user interactions in a financial domain, the framework significantly improves content professionalism and achieves a 32.7% accuracy gain, while eliminating training overhead entirely.

Technology Category

• Machine Learning: Large Multimodal Models (LMMs)
• Application Domains: Humanities & Computational Social Science
• Data Mining & Knowledge Management: Conversational Systems for Recommendation & Retrieval

Application Category

• User Modeling, Personalization and Recommendation: Large Language Models (LLM) for user modeling and recommendation
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Search and Retrieval-Augmented AI: Vertical and domain-specific search

Abstract

Vertical-domain large language models (LLMs) play a crucial role in specialized scenarios such as finance, healthcare, and law; however, their training often relies on large-scale annotated data and substantial computational resources, impeding rapid development and continuous iteration. To address these challenges, we introduce the Collaborative Editable Model (CoEM), which constructs a candidate knowledge pool from user-contributed domain snippets, leverages interactive user-model dialogues combined with user ratings and attribution analysis to pinpoint high-value knowledge fragments, and injects these fragments via in-context prompts for lightweight domain adaptation. With high-value knowledge, the LLM can generate more accurate and domain-specific content. In a financial information scenario, we collect 15k feedback from about 120 users and validate CoEM with user ratings to assess the quality of generated insights, demonstrating significant improvements in domain-specific generation while avoiding the time and compute overhead of traditional fine-tuning workflows.

Problem

• Enables lightweight domain adaptation without large annotated data
• Identifies high-value knowledge via user interactions and ratings
• Improves domain-specific content generation efficiency and accuracy

Innovation

• Constructs candidate knowledge pool from user snippets
• Uses interactive dialogues for high-value knowledge identification
• Injects fragments via prompts for lightweight adaptation