Enhancing Paraphrase Type Generation: The Impact of DPO and RLHF Evaluated with Human-Ranked Data

TL;DR

This paper demonstrates that using human-ranked data and Direct Preference Optimization (DPO) significantly improves paraphrase type generation accuracy and human preference alignment, advancing the reliability of semantic paraphrasing models.

Contribution

It introduces a DPO-based training approach with human-annotated data to enhance paraphrase-type generation accuracy and human preference alignment.

Findings

DPO training increases paraphrase accuracy by 3 percentage points.

Human preference ratings improve by 7 percentage points.

Paraphrase-type detection achieves high F1 scores, e.g., 0.91 for addition/deletion.

Abstract

Paraphrasing re-expresses meaning to enhance applications like text simplification, machine translation, and question-answering. Specific paraphrase types facilitate accurate semantic analysis and robust language models. However, existing paraphrase-type generation methods often misalign with human preferences due to reliance on automated metrics and limited human-annotated training data, obscuring crucial aspects of semantic fidelity and linguistic transformations. This study addresses this gap by leveraging a human-ranked paraphrase-type dataset and integrating Direct Preference Optimization (DPO) to align model outputs directly with human judgments. DPO-based training increases paraphrase-type generation accuracy by 3 percentage points over a supervised baseline and raises human preference ratings by 7 percentage points. A newly created human-annotated dataset supports more rigorous future evaluations. Additionally, a paraphrase-type detection model achieves F1 scores of 0.91 for addition/deletion, 0.78 for same polarity substitution, and 0.70 for punctuation changes. These findings demonstrate that preference data and DPO training produce more reliable, semantically accurate paraphrases, enabling downstream applications such as improved summarization and more robust question-answering. The PTD model surpasses automated metrics and provides a more reliable framework for evaluating paraphrase quality, advancing paraphrase-type research toward richer, user-aligned language generation and establishing a stronger foundation for future evaluations grounded in human-centric criteria.