Fig. 1: The federated machine learning framework of retired battery sorting for recycling.

a Multiple data sources, such as battery manufacturers (Image courtesy Addionics), practical application operators (battery pack in the floor pan of a Tesla. Image courtesy of Tesla), academic research institutions, and third-party platforms, can be data contributors. The battery data are neither exchanged between contributors nor uploaded to the battery recycler. Instead, the data contributors train local models and share model parameters with the battery recycler to build a global model. The proposed Wasserstein-distance voting technique fuses the local models into the global model, which is robust to data imbalance and noise. Battery recyclers can use the jointly-built model for battery sorting, combined with the easy-to-access field testing data. b Our federated machine learning framework encourages collaborators to sharing the data while preserving data privacy as apposed to the traditional data islanding paradigm.