The paper introduces Whole-Piece Training for Symbolic Music Language Models using Full-Horizon Compressed Recurrence (FHCR), which maintains the full temporal horizon of recurrent memory while compressing its key-value representation to fit GPU limits. An evaluation diagnostic, KV-Reset Context Utilization (KRCU), demonstrates that full-horizon models retain long-range context beyond local windows, whereas limiting recurrent memory weakens this dependence. FHCR thus preserves long-range context utilization while significantly reducing recurrent memory cost, enabling efficient whole-piece modeling.
arXiv:2607. 13587v1 Announce Type: cross Abstract: Automatic symbolic music analysis has made substantial progress, yet existing systems are typically designed for a single mode of use, such as full-score prediction, and therefore do not match the broader range of operations that arise in analysis workflows, including partial completion, local correction, and iterative refinement.
The paper explains why GPT‑style language models fail to transfer directly to symbolic music. It argues that success in language comes from tokenization that compresses data by creating a coordinate system where recurring patterns become predictable. For music, the authors propose that tokenization must build a predictively effective, relationally lossless coordinate system—defining Fact–Token and Token–State boundaries—to enable compression without sacrificing contextual freedom. Controlled experiments confirm that proper coordinate construction improves predictive compressibility, whereas mere sequence compaction does not.
The paper proposes the Effectiveness–Losslessness Framework to guide tokenization in domains beyond language, using predictive codelength as a criterion. It introduces two boundaries: the Fact–Token Boundary, where observable structure should be encoded into tokens, and the Token–State Boundary, where context‑dependent relations should remain for model state rather than being pre‑tokenized. Experiments on symbolic music show that making musical time explicit and applying tonal‑frame canonicalization improve predictive performance, while fixed pitch coordinates and reversible BPE can increase predictive code length, indicating that carrier compaction alone does not guarantee better predictions.
TEMPO is a unified model that adds temporally‑grounded capabilities to large audio‑language models, enabling timestamping of events, speakers, and sounds in audio, speech, and music. It introduces a supervised fine‑tuning stage featuring atomic timestamp tokens, a time‑aware projector with sinusoidal encodings, and a distance‑aware Gaussian loss, trained via a synthetic‑to‑real curriculum. Additionally, TEMPO employs reinforcement learning (GRPO) as a refinement step, and achieves state‑of‑the‑art performance on a benchmark of 10K samples across five timestamping tasks, surpassing Audio Flamingo Next and Qwen3‑Omni.
arXiv:2608. 08569v1 Announce Type: new Abstract: Recent advancements in Speech Large Language Models have demonstrated remarkable capabilities in understanding complex audio tasks.
arXiv:2605. 08696v4 Announce Type: replace-cross Abstract: Over the last two decades, language modeling has experienced a shift from the use of predominantly recurrent architectures that process tokens sequentially during training and inference to non-recurrent models that process sequence elements in parallel during training, which results in greater training efficiency and stability at the expense of lower inference throughput.
arXiv:2607. 01849v1 Announce Type: cross Abstract: Musical performance involves executing a set of high-level musical instructions, yet recovering those instructions from the performance is a challenging inverse problem.
arXiv:2604. 19532v3 Announce Type: replace-cross Abstract: Tokenizing music to fit the general framework of language models is a compelling challenge, especially considering the diverse symbolic structures in which music can be represented (e.
MusTBench is a music‑expert‑validated benchmark that evaluates temporal grounding in Large Audio‑Language Models (LALMs) through five temporally grounded question‑answering tasks. The paper also introduces MusT, a four‑stage optimization recipe—music encoder adaptation, LLM adaptation, supervised fine‑tuning, and RL‑based optimization—to improve temporal grounding. Experiments show that current LALMs struggle with precise temporal grounding, while MusT yields significant improvements, highlighting temporal grounding as a key missing capability in these models.
arXiv:2512. 02652v2 Announce Type: replace-cross Abstract: Existing methods for expressive music performance rendering, a conditional generation task that aims to generate a human-like performance from a symbolic score, rely on supervised learning over small labeled datasets, which limits scaling of both data volume and model size, despite the availability of vast unlabeled music, as in vision and language.
arXiv:2607. 11614v1 Announce Type: cross Abstract: Extending the context length of large language models (LLMs) is critical for many real-world applications, yet standard transformers remain constrained by quadratic compute and linear memory scaling.
arXiv:2606. 03169v1 Announce Type: cross Abstract: Recent song generation systems can synthesize realistic audio, yet generating complete songs remains challenging for two reasons.