SAGE-Music: Low-Latency Symbolic Music Generation via Attribute-Specialized Key-Value Head Sharing

• URL: http://arxiv.org/abs/2510.00395v2
• Date: Wed, 15 Oct 2025 02:29:01 GMT
• Title: SAGE-Music: Low-Latency Symbolic Music Generation via Attribute-Specialized Key-Value Head Sharing
• Abstract summary: Low-latency symbolic music generation is essential for real-time and human-AI co-creation.<n>Traditional acceleration techniques such as embedding pooling significantly degrade musical quality.<n>We propose Attribute-Specialized Key-Value Head Sharing (AS-KVHS) to achieve about 30% speedup with only a negligible (about 0.4%) quality drop.
• Score: 38.95508193655578
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Low-latency symbolic music generation is essential for real-time improvisation and human-AI co-creation. Existing transformer-based models, however, face a trade-off between inference speed and musical quality. Traditional acceleration techniques such as embedding pooling significantly degrade quality, while recently proposed Byte Pair Encoding (BPE) methods - though effective on single-track piano data - suffer large performance drops in multi-track settings, as revealed by our analysis. We propose Attribute-Specialized Key-Value Head Sharing (AS-KVHS), adapted to music's structured symbolic representation, achieving about 30% inference speedup with only a negligible (about 0.4%) quality drop in objective evaluations and slight improvements in subjective listening tests. Our main contributions are (1) the first systematic study of BPE's generalizability in multi-track symbolic music, and (2) the introduction of AS-KVHS for low-latency symbolic music generation. Beyond these, we also release SAGE-Music, an open-source benchmark that matches or surpasses state-of-the-art models in generation quality.

Related papers

• FIGMA: Towards FIne-Grained Music retrievAl [65.98380295254817]
  Retrieving music using natural language descriptions has improved with contrastive audio-text models such as CLAP, but current systems remain limited to coarse semantic queries.<n>We show that this limitation stems from the contrastive learning objective itself: despite being trained on long captions, CLAP-based models effectively utilize only the first few tokens.<n>We propose FIGMA, a multi-view contrastive architecture that addresses this limitation by jointly optimizing global audio-text alignment and frame-level, token-wise alignment.
  arXiv  Detail & Related papers  (2026-06-04T18:05:39Z)
• When Noise Lowers The Loss: Rethinking Likelihood-Based Evaluation in Music Large Language Models [4.959996513415697]
  Music large language models (LLMs) need robust methods of evaluating output quality.<n>We show that models encounter systematically corrupted music, undermining its validity as a standalone quality indicator.<n>Our results highlight a new principle: the shape of the loss curve -- rather than its absolute value -- encodes critical information about the quality of the generated content.
  arXiv  Detail & Related papers  (2026-02-02T19:53:12Z)
• BACHI: Boundary-Aware Symbolic Chord Recognition Through Masked Iterative Decoding on Pop and Classical Music [37.085465708242445]
  We introduce POP909-CL, an enhanced version of POP909 dataset with tempo-aligned content and human-corrected labels of chords, beats, keys, and time signatures.<n>We propose BACHI, a symbolic chord recognition model that decomposes the task into different decision steps, namely boundary detection and iterative ranking of chord root, quality, and bass (inversion)<n> Experiments demonstrate that BACHI achieves state-of-the-art chord recognition performance on both classical and pop music benchmarks.
  arXiv  Detail & Related papers  (2025-10-08T00:02:56Z)
• Via Score to Performance: Efficient Human-Controllable Long Song Generation with Bar-Level Symbolic Notation [5.793988399290161]
  Song generation is regarded as the most challenging problem in music AIGC.<n>We present Bar-level AI Composing Helper (BACH), the first model explicitly designed for song generation through human-editable symbolic scores.<n>BACH introduces a tokenization strategy and a symbolic generative procedure tailored to hierarchical song structure.
  arXiv  Detail & Related papers  (2025-08-02T14:58:34Z)
• Scaling Self-Supervised Representation Learning for Symbolic Piano Performance [52.661197827466886]
  We study the capabilities of generative autoregressive transformer models trained on large amounts of symbolic solo-piano transcriptions.<n>We use a comparatively smaller, high-quality subset to finetune models to produce musical continuations, perform symbolic classification tasks, and produce general-purpose contrastive MIDI embeddings.
  arXiv  Detail & Related papers  (2025-06-30T14:00:14Z)
• Refining music sample identification with a self-supervised graph neural network [16.73613870989583]
  We propose a lightweight and scalable encoding architecture employing a Graph Neural Network within a contrastive learning framework.<n>Our model uses only 9% of the trainable parameters compared to the current state-of-the-art system while achieving comparable performance, reaching a mean average precision (mAP) of 44.2%.<n>In addition, because queries in real-world applications are often short in duration, we benchmark our system for short queries using new fine-grained annotations for the Sample100 dataset.
  arXiv  Detail & Related papers  (2025-06-17T16:19:21Z)
• Analyzable Chain-of-Musical-Thought Prompting for High-Fidelity Music Generation [10.643965544581683]
  We introduce MusiCoT, a novel chain-of-thought (CoT) prompting technique tailored for music generation.<n>MusiCoT empowers the AR model to first outline an overall music structure before generating audio tokens.<n>Our experimental results indicate that MusiCoT consistently achieves superior performance across both objective and subjective metrics.
  arXiv  Detail & Related papers  (2025-03-25T12:51:21Z)
• MuseBarControl: Enhancing Fine-Grained Control in Symbolic Music Generation through Pre-Training and Counterfactual Loss [51.85076222868963]
  We introduce a pre-training task designed to link control signals directly with corresponding musical tokens. We then implement a novel counterfactual loss that promotes better alignment between the generated music and the control prompts.
  arXiv  Detail & Related papers  (2024-07-05T08:08:22Z)
• Multi-view MidiVAE: Fusing Track- and Bar-view Representations for Long Multi-track Symbolic Music Generation [50.365392018302416]
  We propose Multi-view MidiVAE, as one of the pioneers in VAE methods that effectively model and generate long multi-track symbolic music. We focus on instrumental characteristics and harmony as well as global and local information about the musical composition by employing a hybrid variational encoding-decoding strategy.
  arXiv  Detail & Related papers  (2024-01-15T08:41:01Z)
• MARBLE: Music Audio Representation Benchmark for Universal Evaluation [79.25065218663458]
  We introduce the Music Audio Representation Benchmark for universaL Evaluation, termed MARBLE. It aims to provide a benchmark for various Music Information Retrieval (MIR) tasks by defining a comprehensive taxonomy with four hierarchy levels, including acoustic, performance, score, and high-level description. We then establish a unified protocol based on 14 tasks on 8 public-available datasets, providing a fair and standard assessment of representations of all open-sourced pre-trained models developed on music recordings as baselines.
  arXiv  Detail & Related papers  (2023-06-18T12:56:46Z)
• Simple and Controllable Music Generation [94.61958781346176]
  MusicGen is a single Language Model (LM) that operates over several streams of compressed discrete music representation, i.e., tokens. Unlike prior work, MusicGen is comprised of a single-stage transformer LM together with efficient token interleaving patterns.
  arXiv  Detail & Related papers  (2023-06-08T15:31:05Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.