Learning Hard Alignments with Variational Inference
This work addresses a specific training bottleneck for hard attention models in sequential tasks like speech recognition, offering a more efficient alternative to existing methods.
The paper tackles the difficulty of training hard attention models due to discrete latent variables by applying variational inference methods (VIMCO and NVIL) with a novel adapted baseline, demonstrating improved performance over REINFORCE on a phoneme recognition task, especially in noisy environments.
There has recently been significant interest in hard attention models for tasks such as object recognition, visual captioning and speech recognition. Hard attention can offer benefits over soft attention such as decreased computational cost, but training hard attention models can be difficult because of the discrete latent variables they introduce. Previous work used REINFORCE and Q-learning to approach these issues, but those methods can provide high-variance gradient estimates and be slow to train. In this paper, we tackle the problem of learning hard attention for a sequential task using variational inference methods, specifically the recently introduced VIMCO and NVIL. Furthermore, we propose a novel baseline that adapts VIMCO to this setting. We demonstrate our method on a phoneme recognition task in clean and noisy environments and show that our method outperforms REINFORCE, with the difference being greater for a more complicated task.