SchenkerLink: Human-in-the-Loop Hierarchical Music Analysis with Uncertainty-Aware Graph Link Prediction

Stephen Ni-Hahn, Jerry Zhu, Chao Péter Yang, Cynthia Rudin, Simon Mak, Yue Jiang
February, 2026

Abstract

SchenkerLink is a human-in-the-loop framework for hierarchical music analysis that leverages uncertainty-aware graph link prediction to model voice-leading and prolongational structures in symbolic music. Unlike prior systems that rely on fixed heuristics or fully automated approaches, SchenkerLink combines the strengths of machine learning and expert feedback to improve interpretability and robustness in music-theoretic analysis. By incorporating uncertainty quantification into the link-prediction process, the system guides analysts toward ambiguous or structurally significant passages, making the analytical process more efficient and musically meaningful. Results demonstrate that SchenkerLink not only improves predictive performance but also supports deeper collaboration between computational models and human music theorists.

Type
Preprint
Publication
Resubmitted to the ACM SIGKDD International Conference on Knowledge Discovery and Data Mining (KDD 2027)
Symbolic Music Analysis Graph Neural Networks Uncertainty Modeling Interpretability Human-in-the-Loop
Chao Péter Yang
Chao Péter Yang
Machine Learning Researcher

My research focuses on LLM alignment, agentic systems, and structure-aware generative modeling. I build preference-aligned language models, reliable multi-agent systems, and symbolic music generation models, and aim to bridge theory and practice to create both scientific and real-world impact.