English(EN)From Prediction to Collaboration: Interactive Symbolic Music Analysis
AI通过新框架推进符号音乐生成与分析 · 跟踪6个来源
作者PulseAugur 编辑部·[7 个来源]·
研究人员正在开发用于符号音乐生成和分析的新AI框架。一种方法MIDI-RAE-JEPA使用带有Swin Transformer V2编码器的自监督学习来捕获分层音乐结构，在情感分类方面优于基线方法。另一篇论文介绍了一个统一的交互式罗马数字分析框架，该框架在预测性能和分析工作流程的响应能力之间取得平衡。此外，BeatEdit提出了一种基于编辑的符号音乐生成方法，将创作视为修订而非从头合成，并且一个神经符号框架通过将语言模型与验证和修复循环集成来增强十二音作曲。
AI
arXiv:2607.14537v1 Announce Type: cross Abstract: Rich internal representations of musical structure are essential for music understanding tasks such as machine-assisted music co-writing, yet self-supervised approaches for symbolic music representation remain underexplored, parti…
arXiv cs.AI
TIER_1English(EN)·Emmanouil Karystinaios, Johannes Hentschel, Markus Neuwirth, Gerhard Widmer·
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 …
Rich internal representations of musical structure are essential for music understanding tasks such as machine-assisted music co-writing, yet self-supervised approaches for symbolic music representation remain underexplored, particularly those that encode the hierarchical multisc…
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 com…
arXiv:2607.11124v1 Announce Type: cross Abstract: Music creation is fundamentally a process of revision. Yet symbolic music generation remains dominated by paradigms that produce complete sequences from scratch, with limited support for selective modification. Edit-based methods …
arXiv:2607.11334v1 Announce Type: new Abstract: Large language models can produce superficially legal twelve-tone scores that collapse into degenerate textures. We introduce a neuro-symbolic harness that wraps a language-model proposer in a generate-verify-repair-trace loop with …
Large language models can produce superficially legal twelve-tone scores that collapse into degenerate textures. We introduce a neuro-symbolic harness that wraps a language-model proposer in a generate-verify-repair-trace loop with symbolic verification. The complete pipeline imp…
