1

PanLUNA: An Efficient and Robust Query-Unified Multimodal Model for Edge Biosignal Intelligence
5.4M-parameter multimodal biosignal foundation model (EEG/ECG/PPG) deployable on a RISC-V microcontroller.
Marija Zelic, Anna Tegon, Yawei Li, Thorir Mar Ingolfsson, Luca Benini
PDF Project DOI
PanLUNA: An Efficient and Robust Query-Unified Multimodal Model for Edge Biosignal Intelligence
LuMamba: Latent Unified Mamba for Electrode Topology-Invariant and Efficient EEG Modeling
Topology-invariant, linear-complexity EEG foundation model; first systematic study of LeJEPA for biosignals.
Danaé Broustail, Anna Tegon, Thorir Mar Ingolfsson, Yawei Li, Luca Benini
PDF Project DOI
LuMamba: Latent Unified Mamba for Electrode Topology-Invariant and Efficient EEG Modeling
LUNA: Efficient and Topology-Agnostic Foundation Model for EEG Signal Analysis
Topology-agnostic EEG foundation model that delivers state-of-the-art performance with 300× fewer FLOPs and 10× lower memory usage.
Berkay Döner, Thorir Mar Ingolfsson, Luca Benini, Yawei Li
PDF Code Project DOI
LUNA: Efficient and Topology-Agnostic Foundation Model for EEG Signal Analysis
FEMBA: Efficient and Scalable EEG Analysis with a Bidirectional Mamba Foundation Model
Bidirectional Mamba EEG foundation model that matches transformer accuracy while scaling linearly with sequence length.
Anna Tegon, Thorir Mar Ingolfsson, Xiaying Wang, Luca Benini, Yawei Li
PDF Code Project DOI
FEMBA: Efficient and Scalable EEG Analysis with a Bidirectional Mamba Foundation Model
Finetuning and Quantization of EEG-Based Foundational BioSignal Models on ECG and PPG Data for Blood Pressure Estimation
Transfers EEG foundation models to ECG/PPG for BP estimation, reaching 1.57 mmHg MAE with 3.5× model compression via INT8 quantization.
Bálint Tóth, Dominik Senti, Thorir Mar Ingolfsson, Jeffrey Zweidler, Alexandre Elsig, Luca Benini, Yawei Li
PDF Project DOI
Train-On-Request: An On-Device Continual Learning Workflow for Adaptive Real-World Brain-Machine Interfaces
On-device continual learning for wearable BMIs that reaches 92% accuracy with 1.6 min recalibration and 1 mJ training steps.
Lan Mei, Cristian Cioflan, Thorir Mar Ingolfsson, Victor Kartsch, Andrea Cossettini, Xiaying Wang, Luca Benini
PDF Code DOI
VowelNet: Enhancing Communication with Wearable EEG-Based Vowel Imagery
Lightweight imagined-speech decoder that runs on GAP9 at 17 mW using only eight EEG channels.
Thorir Mar Ingolfsson, Victor Javier Kartsch Morinigo, Andrea Cossettini, Xiaying Wang, Luca Benini
PDF PDF (ETH Research Collection)
BISeizuRe: BERT-Inspired Seizure Data Representation to Improve Epilepsy Monitoring
BERT-inspired EEG encoder that cuts seizure-detection false positives to 0.23 FP/h after subject-specific tuning.
Luca Benfenati, Thorir Mar Ingolfsson, Andrea Cossettini, Daniele Jahier Pagliari, Alessio Burrello, Luca Benini
PDF Dataset Project Preprint (arXiv) TUH EEG Corpus
12 mJ per Class On-Device Online Few-Shot Class-Incremental Learning
On-device online few-shot class-incremental learning at 12 mJ per class for extreme-edge systems.
Yoga Esa Wibowo, Cristian Cioflan, Thorir Mar Ingolfsson, Michael Hersche, Leo Zhao, Abbas Rahimi, Luca Benini
PDF DOI
Skilog: A Smart Sensor System for Performance Analysis and Biofeedback in Ski Jumping
Wearable ski-jumping coaching system with 92.7% phase accuracy running on GAP9 with multi-day autonomy.
Lukas Schulthess, Thorir Mar Ingolfsson, Marc Nölke, Michele Magno, Luca Benini, Christoph Leitner
PDF Cite IEEE Xplore
Skilog: A Smart Sensor System for Performance Analysis and Biofeedback in Ski Jumping