Evaluating the Effectiveness of Direct Preference Optimization for Personalizing German Automatic Text Simplifications for Persons with Intellectual Disabilities

2025-07-02

AI Summary

Existing German automatic text simplification (ATS) systems fail to model the preferences of users with intellectual disabilities, resulting in insufficient personalization. This paper introduces direct preference optimization (DPO) for the first time into ATS personalization tailored to this population, establishing an end-to-end pipeline encompassing real-user preference data collection, large language model selection, supervised fine-tuning (SFT), and DPO-based alignment. Crucially, it centers target users’ active participation in both data generation and feedback, enabling precise alignment between model outputs and readability requirements. Experiments demonstrate that DPO significantly improves the acceptability and utility of simplified texts, validating the effectiveness of human-preference-aligned personalization. The work provides a reproducible methodological framework and empirical evidence for inclusive, accessible AI systems.

Technology Category

• Humans and AI: Learning Human Values and Preferences
• Machine Learning: Learning Preferences or Rankings
• Knowledge Representation and Reasoning: Preferences

Application Category

• User Modeling, Personalization and Recommendation: User privacy protection in personalized systems
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Economics, Online Markets and Human Computation: Data quality aspects of human-annotated datasets

Abstract

Automatic text simplification (ATS) aims to enhance language accessibility for various target groups, particularly persons with intellectual disabilities. Recent advancements in generative AI, especially large language models (LLMs), have substantially improved the quality of machine-generated text simplifications, thereby mitigating information barriers for the target group. However, existing LLM-based ATS systems do not incorporate preference feedback on text simplifications during training, resulting in a lack of personalization tailored to the specific needs of target group representatives. In this work, we extend the standard supervised fine-tuning (SFT) approach for adapting LLM-based ATS models by leveraging a computationally efficient LLM alignment technique -- direct preference optimization (DPO). Specifically, we post-train LLM-based ATS models using human feedback collected from persons with intellectual disabilities, reflecting their preferences on paired text simplifications generated by mainstream LLMs. Furthermore, we propose a pipeline for developing personalized LLM-based ATS systems, encompassing data collection, model selection, SFT and DPO post-training, and evaluation. Our findings underscore the necessity of active participation of target group persons in designing personalized AI accessibility solutions aligned with human expectations. This work represents a step towards personalizing inclusive AI systems at the target-group level, incorporating insights not only from text simplification experts but also from target group persons themselves.

Problem

• Personalizing German text simplifications for intellectual disabilities
• Incorporating human feedback into LLM-based simplification systems
• Developing a pipeline for personalized ATS using DPO and SFT

Innovation

• Uses Direct Preference Optimization for personalization
• Incorporates human feedback from disabled individuals
• Proposes pipeline for personalized ATS system development