Abstract:
AI healthcare applications use sensitive EHRs that are rarely labeled and distributed across symbiont institutions. Such data makes machine learning model training difficult. To address these challenges, we propose dynamic neural graphs-based federated learning framework. The framework federates Reptile, a model-agnostic meta-learning (MAML) algorithm. This paper proposes a dynamic neural graph learning (NGL) algorithm to include unlabeled examples in supervised training, unlike MAML algorithms. Dynamic NGL computes a meta-learning update by supervised learning on a labelled training example and metric learning on its labelled or unlabelled neighborhood. Local graphs built over training examples establish a labelled example’s neighbourhood dynamically. Each local graph is created by comparing the model’s current embeddings. This semi-supervised framework uses neighborhood metric learning. The publicly available MIMIC-III dataset shows that the proposed framework works for single and multi-task settings under data decentralisation constraints and limited supervision.
Note: Please discuss with our team before submitting this abstract to the college. This Abstract or Synopsis varies based on student project requirements.
Did you like this final year project?
To download this project Code with thesis report and project training... Click Here
