Federated Learning for Emoji Prediction in a Mobile Keyboard
This work demonstrates the feasibility of using federated learning for production-quality natural language understanding tasks while keeping user data on devices, addressing privacy concerns in mobile applications.
The authors tackled the problem of predicting emoji from text typed on a mobile keyboard using a word-level recurrent neural network, achieving better performance with a federated learning model compared to a server-trained model.
We show that a word-level recurrent neural network can predict emoji from text typed on a mobile keyboard. We demonstrate the usefulness of transfer learning for predicting emoji by pretraining the model using a language modeling task. We also propose mechanisms to trigger emoji and tune the diversity of candidates. The model is trained using a distributed on-device learning framework called federated learning. The federated model is shown to achieve better performance than a server-trained model. This work demonstrates the feasibility of using federated learning to train production-quality models for natural language understanding tasks while keeping users' data on their devices.