A Trust-Network-Based Federated Learning Framework for Multi-Center Aging Clock Prediction

The paper introduces TNFL, a trust‑network‑based federated learning framework designed for multi‑center aging clock prediction. TNFL propagates models along directed trust relations without centralized aggregation, combining an age‑aware mixture‑of‑experts model with generative replay to mitigate forgetting and drift. Experiments on multiple molecular datasets demonstrate effective aging‑clock prediction with limited local data, interpretable age‑dependent patterns, and stable performance across interaction orders, while revealing coordinated higher‑order protein subnetworks linked to aging.

Machine-generated by The Flow from the publisher's headline and feed description — not written or checked by a human. The full article lives at arXiv AI.

• safety