Fast Adaptation in Generative Models with Generative Matching Networks
This work addresses the bottleneck of fast adaptation in generative models for tasks like one-shot learning, offering a novel approach that could benefit applications requiring quick learning from limited data, though it appears incremental as it builds on matching networks from discriminative tasks.
The paper tackles the problem of slow adaptation and poor generalization from few examples in deep generative models by introducing Generative Matching Networks, which condition on additional input data to instantly learn new concepts, achieving significant improvements in predictive performance on the Omniglot dataset as more data becomes available and outperforming existing state-of-the-art conditional generative models.
Despite recent advances, the remaining bottlenecks in deep generative models are necessity of extensive training and difficulties with generalization from small number of training examples. We develop a new generative model called Generative Matching Network which is inspired by the recently proposed matching networks for one-shot learning in discriminative tasks. By conditioning on the additional input dataset, our model can instantly learn new concepts that were not available in the training data but conform to a similar generative process. The proposed framework does not explicitly restrict diversity of the conditioning data and also does not require an extensive inference procedure for training or adaptation. Our experiments on the Omniglot dataset demonstrate that Generative Matching Networks significantly improve predictive performance on the fly as more additional data is available and outperform existing state of the art conditional generative models.