Pianist Transformer advances expressive music generation with self-supervised learning

Researchers have developed Pianist Transformer, a novel approach to generating expressive piano performances from symbolic music scores. This method utilizes large-scale self-supervised learning on over 10 billion tokens of unlabeled MIDI data, a significant advancement over previous supervised methods that were limited by small datasets. The model incorporates an efficient asymmetric Transformer architecture with note-level compression to enhance training and inference efficiency, and it offers an editable workflow for integration into music production. The project aims to pave a scalable path toward synthesizing human-like musical performances. AI

IMPACT Enables more scalable and efficient generation of expressive musical performances, potentially impacting music production and AI-driven creativity.

RANK_REASON The cluster contains a research paper detailing a new model and methodology for music generation. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.AI →

• alphaXiv
• arXiv
• CatalyzeX
• DagsHub
• Gotit.pub
• Hong-Jie You
• Hugging Face
• MIDI
• Pianist Transformer
• ScienceCast

• paper
• model release

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →