Learning Longer-term Dependencies in RNNs with Auxiliary Losses
This addresses a fundamental problem in sequence modeling for researchers and practitioners, offering a resource-efficient solution, though it is incremental as it builds on existing RNN and auxiliary loss techniques.
The paper tackles the challenge of capturing long-term dependencies in RNNs by proposing a method that adds an unsupervised auxiliary loss to the original objective, enabling truncated backpropagation for sequences up to 16,000 steps and improving performance on tasks like pixel-by-pixel image classification and document classification.
Despite recent advances in training recurrent neural networks (RNNs), capturing long-term dependencies in sequences remains a fundamental challenge. Most approaches use backpropagation through time (BPTT), which is difficult to scale to very long sequences. This paper proposes a simple method that improves the ability to capture long term dependencies in RNNs by adding an unsupervised auxiliary loss to the original objective. This auxiliary loss forces RNNs to either reconstruct previous events or predict next events in a sequence, making truncated backpropagation feasible for long sequences and also improving full BPTT. We evaluate our method on a variety of settings, including pixel-by-pixel image classification with sequence lengths up to 16\,000, and a real document classification benchmark. Our results highlight good performance and resource efficiency of this approach over competitive baselines, including other recurrent models and a comparable sized Transformer. Further analyses reveal beneficial effects of the auxiliary loss on optimization and regularization, as well as extreme cases where there is little to no backpropagation.