LGAIMar 11, 2025

Preference-Based Alignment of Discrete Diffusion Models

arXiv:2503.08295v210 citationsh-index: 44
Originality Incremental advance
AI Analysis

This work addresses a practical problem for researchers and practitioners using discrete diffusion models in domains like language modeling and protein sequence generation, though it appears incremental as it adapts an existing method to a new model type.

The paper tackles the challenge of aligning discrete diffusion models with task-specific preferences when explicit reward functions are unavailable, by introducing Discrete Diffusion DPO (D2-DPO), which adapts Direct Preference Optimization to discrete diffusion models and demonstrates effective alignment on a structured binary sequence generation task.

Diffusion models have achieved state-of-the-art performance across multiple domains, with recent advancements extending their applicability to discrete data. However, aligning discrete diffusion models with task-specific preferences remains challenging, particularly in scenarios where explicit reward functions are unavailable. In this work, we introduce Discrete Diffusion DPO (D2-DPO), the first adaptation of Direct Preference Optimization (DPO) to discrete diffusion models formulated as continuous-time Markov chains. Our approach derives a novel loss function that directly fine-tunes the generative process using preference data while preserving fidelity to a reference distribution. We validate D2-DPO on a structured binary sequence generation task, demonstrating that the method effectively aligns model outputs with preferences while maintaining structural validity. Our results highlight that D2-DPO enables controlled fine-tuning without requiring explicit reward models, making it a practical alternative to reinforcement learning-based approaches. Future research will explore extending D2-DPO to more complex generative tasks, including language modeling and protein sequence generation, as well as investigating alternative noise schedules, such as uniform noising, to enhance flexibility across different applications.

Foundations

The foundational work for this paper's niche, ranked by how specifically the neighbourhood builds on it — not by global fame.

Your Notes