比较三类模型生成巴赫风格钢琴乐谱，发现自回归模型最连贯，向量量化提升结构质量。
Generative Modeling of Bach-Style Symbolic Music: A Comparative Study of Autoregressive, Latent-Variable, and Adversarial Approaches
- 用注意力LSTM、变分自编码器和对抗网络生成巴赫风格乐谱
- 自回归模型生成的乐句最符合音乐逻辑，向量量化缓解了后验崩溃
- 适合对音乐生成机制或模型对比感兴趣的读者
我们基于同一MIDI语料库，对比了三类模型在生成巴赫风格钢琴乐谱中的表现：带注意力的自回归LSTM、包含循环VAE和向量量化VAE的潜在变量模型，以及生成对抗网络。实验评估其建模多声部音符序列的能力、学习有用潜在表示的能力，以及生成风格一致作品的能力。结果表明，带注意力的自回归LSTM生成的样本最具音乐连贯性；向量量化有效缓解了后验崩溃问题，使输出更具结构性，优于传统循环VAE；而对抗方法虽能捕捉局部音高模式，但训练困难且对巴赫风格泛化能力较差。这些结果揭示了自回归、潜在变量与对抗方法在符号音乐生成中的优劣与失效模式。
原文摘要 · Abstract (English)
We study generative modeling of Bach-style symbolic piano music using a shared MIDI corpus and three model families: autoregressive LSTMs with attention, latent-variable models including recurrent VAEs and vector-quantized VAEs, and generative adversarial networks. We compare their ability to model polyphonic note sequences, learn useful latent representations, and generate stylistically coherent compositions. Our experiments show that the autoregressive LSTM with attention produces the most musically coherent samples, while vector quantization helps mitigate posterior collapse and yields more structured outputs than conventional recurrent VAEs. The adversarial approach captures local pitch patterns but remains difficult to train and generalizes less reliably to Bach's style. These results highlight the relative strengths and failure modes of autoregressive, latent-variable, and adversarial approaches for symbolic music generation.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。