AI Paper Atlas

Research map / Architectures and efficiency / RNNs and normalization

PhoneMD: Learning to Diagnose Parkinson’s Disease from Smartphone Data

Patrick Schwab, Walter Karlen

AAAI 2019 · Architectures and efficiency › RNNs and normalization · approach: Brain-inspired recurrent architectures · 51 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
Aggregate multiple test predictions with EAM and add age and genderPhoneMD · mPower test setAUC: 0.74 -> 0.85Improved
Use EAM instead of mean aggregationNeural-network test predictions · mPower test setAUC: 0.64 -> 0.71Improved
Combine neural-network and expert-feature test modelsEAM · mPower test setAUC: 0.71 -> 0.70Mixed
Add age and gender to the EAM (ablation)EAM (Both) · mPower test setAUC: 0.70 -> 0.85Improved
Train one end-to-end network to assess tests and aggregate them jointly (ablation)PhoneMD · mPower test setAUC: 0.50 (0.50, 0.56)Hurt
Add timing information to the EAM metadata (ablation)EAM · mPower test setpredictive performanceNo clear effect
Add a segment-level neural attention mechanism to the voice CNN (ablation)Voice CNN · Voicepredictive performanceHurt
Downsample walking, voice, and tapping signals (ablation)Specialised test models · Walking, voice, and tappingpredictive performanceNo clear effect
Use the CNN rather than expert features for tappingTapping test model · TappingAUC: 0.62 vs 0.59Hurt
Use expert features rather than the RNN for memoryMemory test model · MemoryAUC: 0.52 vs 0.65Hurt

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