Temporal FiLM: Capturing Long-Range Sequence Dependencies with Feature-Wise Modulations
This addresses a key bottleneck in deep learning for sequential data, offering a novel hybrid method that is incremental in combining existing techniques.
The paper tackles the problem of capturing long-range dependencies in sequential data by proposing Temporal Feature-Wise Linear Modulation (TFiLM), which uses a recurrent neural network to modulate convolutional activations, resulting in significant improvements in learning speed and accuracy on tasks like text classification and audio super-resolution.
Learning representations that accurately capture long-range dependencies in sequential inputs -- including text, audio, and genomic data -- is a key problem in deep learning. Feed-forward convolutional models capture only feature interactions within finite receptive fields while recurrent architectures can be slow and difficult to train due to vanishing gradients. Here, we propose Temporal Feature-Wise Linear Modulation (TFiLM) -- a novel architectural component inspired by adaptive batch normalization and its extensions -- that uses a recurrent neural network to alter the activations of a convolutional model. This approach expands the receptive field of convolutional sequence models with minimal computational overhead. Empirically, we find that TFiLM significantly improves the learning speed and accuracy of feed-forward neural networks on a range of generative and discriminative learning tasks, including text classification and audio super-resolution