A Decomposed Retrieval-Edit-Rerank Framework for Chord Generation

2026-05-08

AI Summary

This work addresses the challenge of balancing stylistic diversity and music-theoretic feasibility in chord generation, a limitation of existing end-to-end approaches that tightly couple candidate generation with constraint satisfaction. To overcome this, the authors propose a Retrieval–Editing–Reranking (RER) framework that systematically decouples the task into three interpretable and controllable stages: diverse candidates are first retrieved to capture stylistic variation; rule-based or learned editing ensures harmonic validity; and a learning-driven reranker integrates soft user preferences to select final outputs. Although end-to-end trainable, this functionally modular pipeline significantly enhances controllability and transparency in navigating the trade-off between creativity and constraint adherence. Experiments demonstrate consistent superiority over current end-to-end baselines in both objective metrics and subjective evaluations, while ablation studies confirm the complementary roles of each stage in creative exploration and theoretical compliance.

Abstract

Chord generation is an inherently constrained creative task that requires balancing stylistic diversity with music-theoretic feasibility. Existing approaches typically entangle candidate generation and constraint enforcement within a single model, making the diversity-feasibility trade-off difficult to control and interpret. In this work, we approach chord generation from a system-level perspective, introducing a Retrieval-Edit-Rerank (RER) framework that decomposes the task into three explicit stages: i) retrieval, which defines a stylistically plausible candidate space; ii) editing, which enforces music-theoretic feasibility through minimal modifications; and iii) reranking, which resolves soft preferences among feasible candidates. This separation provides a controllable pipeline, where each component addresses a distinct aspect of the generation process, thereby enhancing both the interpretability and adjustability of the output chords. Through objective metrics and subjective evaluation, our decomposed system outperforms all end-to-end chord generation baselines in balancing chord diversity and music-theoretic feasibility. Ablation studies further confirm the complementary roles of each stage in creative exploration and constraint satisfaction.

Problem

• chord generation
• stylistic diversity
• music-theoretic feasibility
• constrained creative task
• diversity-feasibility trade-off

Innovation

• Retrieval-Edit-Rerank
• chord generation
• music-theoretic feasibility
• stylistic diversity
• decomposed framework