Research Areas
The BCML Lab has covered a wide span of AI-driven healthcare and neuromorphic computing topics.
Machine Learning / Deep Learning
Signal understanding and perception pipelines that combine CNNs, Transformers, and multimodal fusion to support sleep staging, activity recognition, and embedded inference.
- Sleep/ECG interpretation for wearable healthcare
- Multimodal human-action and sensor fusion models
- Edge-ready architectures with knowledge distillation
Physical AI & Sim-to-Real
AI systems that perceive and act in the physical world, using physics-based simulation to bridge the gap to real environments.
- Physics-based simulation and synthetic data generation
- Sim-to-real gap analysis
- Embodied control for robotic manipulators on embedded and neuromorphic hardware
Reinforcement Learning
Reward-driven agents that learn which biosignal features and decisions matter most, from ECG-based authentication to EMG-driven device control.
- RL-based feature selection over ECG and EMG descriptors
- Reward shaping and curriculum design for sample-efficient learning
- Decision-making policies for biometric and assistive applications
Spiking Neural Networks
Third-generation neural models that exchange 1-bit spikes for low-power, explainable biosignal analytics on neuromorphic hardware.
- Bio-plausible neuron/synapse modeling
- ANN-to-SNN conversion and direct spike training
- ECG/EEG/PPG analytics with interpretable spikes
Computational Neuroscience
Clinician-partnered modeling of cognitive and physiological processes for remote assignments, tele-neuroscience coursework, and virtual assessments.
- Attention, memory, and sensory encoding models
- Remote cognitive assessment and coursework support
- EEG/ECG biomarkers for fatigue and motor intent
Digital Healthcare
AI platforms for preventive care, remote monitoring, and diagnostic decision support, anchored by our Nailfold Capillaroscopy effort.
- Nailfold Capillaroscopy image analysis and vessel enhancement
- Privacy-aware patient monitoring and digital therapeutics
- Wearable signal fusion for cardiometabolic risk tracking
Anomaly Detection
Reliability research that keeps hospitals, wearables, and factories aware of rare events through multimodal outlier detection and lightweight edge models.
- Multisensor anomaly scoring for smart hospitals
- Edge-ready monitoring for wearable devices
- Self-supervised representations for rare-event alerts