Research
My research spans learning from neural signals, connecting brain activity with visual concepts, and evaluating intelligent systems for neural sensing.
Neural State Prediction: Obstructing Shortcut Learning in EEG Foundation Models
Investigating latent prediction objectives that obstruct positional shortcuts and local interpolation, with the goal of learning EEG representations that depend on meaningful signal content.
Self-supervised learning · Representation learning
ProtoEEG: Structured Prototype Alignment for EEG Visual Decoding
A shared learnable visual codebook aligns EEG and visual representations through structured prototype assignments for visual decoding.
Multimodal learning · EEG visual decoding
Signal-grounded reasoning over EEG
Developing an evaluation framework to study whether language model agents reason from EEG evidence, using executable tasks, localized signal evidence, and controlled comparisons of tools and agents.
LLM agents · Evidence-based evaluation
Benchmarking EEG visual decoding
Working toward comparable evaluation of EEG-based visual recognition, retrieval, and reconstruction, with attention to data splits, preprocessing, model provenance, and evaluation protocols.
Benchmarking · Reproducibility
An End-to-End Earbud EEG Platform for Affective and Cognitive Load Assessment in Naturalistic IoT Environments
Studying ear-worn EEG as a practical interface for affective and cognitive load assessment, connecting wearable acquisition hardware with signal representation learning.
Ear-EEG · Affective computing · IoT
Interested in these directions? Get in touch →