An Online Algorithm for Learning over Constrained Latent Representations using Multiple Views

We introduce an online framework for discriminative learning problems over hidden structures, where we learn both the latent structure and the classifier for a supervised learning task. Previous work on leveraging latent representations for discriminative learners has used batch algorithms that require multiple passes though the entire training data. Instead, we propose an online algorithm that efficiently jointly learns the latent structures and the classifier. We further extend this to include multiple views on the latent structures with different representations. Our proposed online algorithm with multiple views significantly outperforms batch learning for latent representations with a single view on a grammaticality prediction task.