SparseMAP: Differentiable Sparse Structured Inference
This addresses the challenge of efficient and interpretable structured inference for machine learning applications like natural language processing, though it is incremental as it builds on existing MAP and marginal inference methods.
The paper tackles the combinatorial search problem in structured prediction by introducing SparseMAP, a method for sparse structured inference that automatically selects a few global structures, resulting in competitive accuracy in dependency parsing and natural language inference tasks.
Structured prediction requires searching over a combinatorial number of structures. To tackle it, we introduce SparseMAP: a new method for sparse structured inference, and its natural loss function. SparseMAP automatically selects only a few global structures: it is situated between MAP inference, which picks a single structure, and marginal inference, which assigns probability mass to all structures, including implausible ones. Importantly, SparseMAP can be computed using only calls to a MAP oracle, making it applicable to problems with intractable marginal inference, e.g., linear assignment. Sparsity makes gradient backpropagation efficient regardless of the structure, enabling us to augment deep neural networks with generic and sparse structured hidden layers. Experiments in dependency parsing and natural language inference reveal competitive accuracy, improved interpretability, and the ability to capture natural language ambiguities, which is attractive for pipeline systems.