Neural Network Research, Lee Lab
Neural Network Research, Lee Lab
Embedded Files
Schematic diagram of the hierarchical neural network employed
I collaborated with Prof. Yoonsang Lee at Dartmouth College to develop a Hierarchical Neural Network (HiNN) for predicting the severity of Parkinson’s Disease (PD) symptoms using speech data.
Our HiNN model analyzed speech signal features from patients at various stages of the disease to estimate motor-UPDRS scores, a key clinical measure of PD symptom progression.
The HiNN structure is unique, splitting the learning process into two levels:
Level 1 - Simple Patterns: Captures broad, low-complexity patterns in the data.
Level 2 - Complex Details: Refines predictions by focusing on high-complexity patterns missed by the first level.
This hierarchical approach processes data in stages, improving efficiency and reducing the risk of overfitting.
Google Sites
Report abuse