AI & ML
Multi-Phonation Graph Learning with Self-Supervised Speech Embeddings for ALS Detection and Progression Prediction
arXiv cs.SD0 views
arXiv:2607.25284v2 Announce Type: replace-cross Abstract: Amyotrophic lateral sclerosis (ALS) progressively impairs speech motor control, making acoustic analysis a promising biomarker for severity and progression estimation. We propose a subject-level graph framework that aggregates multiple phonation recordings into a unique k-nearest-neighbor graph built from pretrained SSL embeddings of 2s segments. We compare four SSL front-ends (wav2vec 2.0, HuBERT, data2vec-audio, and UniSpeech-SAT) and five graph neural networks (GCN, residual GCN, GAT, GraphSAGE, and GIN) on the SAND dataset tasks (339 participants: 205 ALS, 134 control): 5-class dysarthria severity and 4-class ALSFRS-R progression prediction. On the official validation set, the best configuration (HuBERT+GIN) achieves macro-F$_1$ of 0.73 for Task 1 and 0.69 for Task 2, outperforming SAND validation baselines (0.61 and 0.58). These results highlight the potential of combining GNNs with pretrained cross-lingual speech representations for low-resource ALS detection and progression monitoring.
Source：https://arxiv.org/abs/2607.25284PreviousMinimizing Modality Gap from the Input Side: Your Speech LLM Can Be a Prosody-Aware Text LLMNextTamilEOT: A Dataset and Model for Semantic End-of-Turn Detection in Tamil Telephone Speech
Related posts
Do Top-Ranked Local Businesses Get Recommended by ChatGPT? The November 9, 2026 Re-Check: 0 of 3
0dev.to AI
How to Buy ChatGPT Plus Cheap Without Getting Burned — Top-Ups, Pre-Made Accounts, and CDKs Explained
0dev.to AI
Implementing Capacity Providers in ECS on EC2 for Fully Automated Batch Processing Scaling
0dev.to AWS
Self-heal broken locators with a 2B model on your own GPU
0dev.to Python
AWS Config Costs Were Too High—I Cut Them in Half With One Configuration Change
0dev.to DevOps
Nominal, which builds software to make hardware manufacturing more efficient, raised $240M at a $3B valuation, up from $1B after raising $80M in March (Paayal Zaveri/Bloomberg)
0Techmeme