AI Paper Atlas

Research map / Architectures and efficiency / Attention and transformers

Pianist Transformer: Towards Expressive Piano Performance Rendering via Scalable Self-Supervised Pre-Training

Hong-Jie You, Jie-Jing Shao, Xiao-Wen Yang, Lin-Han Jia, Lan-Zhe Guo, Yu-Feng Li

ICML 2026 · Architectures and efficiency › Attention and transformers · approach: Transformer expressivity and circuits · 0 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add large-scale self-supervised pre-training (ablation)Pianist Transformer · ASAP testoverall intersection area: 0.6032 -> 0.8501Improved
add large-scale self-supervised pre-training (ablation)Pianist Transformer · ASAP testoverall JS divergence: 0.4103 -> 0.1634Improved
add pre-training (ablation)Pianist Transformer · ASAP test (Velocity)JS divergence: 0.5363 -> 0.1805Improved
add pre-training (ablation)Pianist Transformer · ASAP test (Duration)JS divergence: 0.5399 -> 0.1879Improved
add pre-training (ablation)Pianist Transformer · ASAP test (IOI)JS divergence: 0.2789 -> 0.1740Improved
add pre-training (ablation)Pianist Transformer · ASAP test (Pedal)JS divergence: 0.2860 -> 0.1111Improved
use asymmetric 10-2 encoder-decoder instead of symmetric 6-6 (ablation)Pianist Transformer · ASAP testoverall JS divergence: 0.174 -> 0.163Improved
use asymmetric 10-2 instead of symmetric 6-6 (ablation)Pianist Transformer · efficiencyCPU inference speed: 1.0x -> 2.1xImproved
use asymmetric 10-2 instead of symmetric 6-6 (ablation)Pianist Transformer · efficiencytraining VRAM: 1.0x -> 0.6xImproved
use symmetric 6-6 instead of 2-layer decoder (ablation)Pianist Transformer 135M · pre-training validationloss: 1.260 -> 1.230Improved
scale pre-training data from 1B to 10B tokens (ablation)Pianist Transformer 135M · pre-training validationloss: 1.199 -> 1.195No clear effect
increase model size from 65M to 135M (ablation)Pianist Transformer · pre-training validationlossNo clear effect

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy