Modular Retrieval-Augmented Generalization for Human Action Recognition

2026-05-12T08:51:40Z590314943d3e4d878d377789db6f9113e028017fbbe6525a6618361f6d440942
EEG-denoisingIMURF-inverse-designTMS-EEGUAVarrhythmia-classificationchannel-gain-mappingdata-collectiondomain-adaptationhuman-activity-recognitionmachine-learningmedical-devicesneural-posterior-estimationopen-sourceprivacyradar-soundingremote-sensingretrieval-augmentedsmartphone-EEGstructural-health-monitoringtelemedicinetrajectory-planningtransformerwearable-ECGwireless-security

What happened

This silver feed bundles ten newly announced papers (arXiv, 12 May 2026) across signal processing, machine learning, sensing, and wireless systems: (1) MoRA — a retrieval-augmented module for IMU-based human activity recognition (open-source) improving performance via uncertainty-adaptive fusion; (2) large multicenter study validating smartphone-based 27-lead EEG acquisition with remote interpretation in Kenya (3,036 sessions) showing high feasibility but potential operational failure modes (high-impedance, short recordings); (3) outage-aware spatio-temporal graph reconstruction for high-rise-

Why it matters

A reviewed impact interpretation has not been published for this record.

Evidence and limitations

Source ID
arxiv_eess_sp
Record identifier
590314943d3e4d878d377789db6f9113e028017fbbe6525a6618361f6d440942
Enrichment time
2026-05-12T08:51:40Z
AI-assisted enrichment
Yes

This record may overlap with other records. Its enrichment can be incomplete or wrong, and machine assistance was used. Validate consequential decisions against the linked source and your own environment.