Transfer learning for ensembles: reducing computation time and keeping the diversity
This work addresses computational efficiency for practitioners using ensemble transfer learning, though it is incremental as it builds on existing transfer learning and ensemble methods.
The paper tackles the high computational cost and overfitting risk in transferring ensembles of deep neural networks by introducing a two-step method: shifting encoder weights with a single vector and fine-tuning each model individually, achieving competitive results with reduced training time while maintaining ensemble diversity.
Transferring a deep neural network trained on one problem to another requires only a small amount of data and little additional computation time. The same behaviour holds for ensembles of deep learning models typically superior to a single model. However, a transfer of deep neural networks ensemble demands relatively high computational expenses. The probability of overfitting also increases. Our approach for the transfer learning of ensembles consists of two steps: (a) shifting weights of encoders of all models in the ensemble by a single shift vector and (b) doing a tiny fine-tuning for each individual model afterwards. This strategy leads to a speed-up of the training process and gives an opportunity to add models to an ensemble with significantly reduced training time using the shift vector. We compare different strategies by computation time, the accuracy of an ensemble, uncertainty estimation and disagreement and conclude that our approach gives competitive results using the same computation complexity in comparison with the traditional approach. Also, our method keeps the ensemble's models' diversity higher.