AI Paper Atlas

Research map / Image generation / Autoregressive and latent generation

SIGMA-PPG: Statistical-prior Informed Generative Masking Architecture for PPG Foundation Model

Zongheng Guo, Tao Chen, Yang Jiao, Yi Pan, Xiao Hu, Manuela Ferrario

ICML 2026 · Image generation › Autoregressive and latent generation · approach: Generative latent space tooling · 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 semantic consistency loss L_Con to the tokenizer (ablation)SIGMA-PPG (30M) · Systolic BP, linear probingMAE: 15.62 -> 13.07Improved
add semantic consistency loss L_Con to the tokenizer (ablation)SIGMA-PPG (30M) · Heart rate, linear probingMAE: 9.21 -> 5.83Improved
add semantic consistency loss L_Con to the tokenizer (ablation)SIGMA-PPG · full fine-tuning, all downstream tasksMAE and AUCImproved
replace random masking with prior-guided adversarial masking (ablation)SIGMA-PPG · regression and classification downstream tasksMAE and AUCImproved
use unconstrained adversarial Teacher (no statistical prior) instead of prior-guided (ablation)SIGMA-PPG · pre-training maskingrepresentation qualityHurt
use static prior-based masking instead of adaptive Teacher (ablation)SIGMA-PPG · pre-training maskingrepresentation qualityHurt
use standard random masking (ablation)SIGMA-PPG · pre-training maskingrepresentation qualityHurt
use generative masked pretraining vs best contrastive baselineSIGMA-PPG (30M) · SpO2, BIDMC, full fine-tuningMAE: 1.265 -> 0.1457Improved
scale SIGMA-PPG from 5M to 30M parameters (ablation)SIGMA-PPG · full fine-tuning averageAUC: 0.8596 -> 0.9082Improved
use SIGMA-PPG instead of Pulse-PPG under linear probing on classificationSIGMA-PPG (30M) · linear probing classification averageAUC: 0.8820 -> 0.8510Hurt
use SIGMA-PPG instead of Pulse-PPG under linear probing on regressionSIGMA-PPG (30M) · linear probing regression averageMAE: 6.781 -> 6.636Improved
use SIGMA-PPG (30M) instead of Pulse-PPG for computeSIGMA-PPG (30M) · Computational costFLOPs: 27.11 -> 2.56 GMacImproved

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