Graph Structured Prediction Energy Networks
This addresses the limitation in classical structured prediction methods that struggle with modeling high-order correlations or explicit known correlations, potentially benefiting researchers and practitioners in NLP and computer vision, though it appears incremental as it builds on existing energy network concepts.
The paper tackles the problem of joint inference over multiple variables in structured prediction by introducing Graph Structured Prediction Energy Networks, which model both explicit local and implicit higher-order correlations while maintaining tractable inference, and demonstrates its utility on tasks in natural language processing and computer vision.
For joint inference over multiple variables, a variety of structured prediction techniques have been developed to model correlations among variables and thereby improve predictions. However, many classical approaches suffer from one of two primary drawbacks: they either lack the ability to model high-order correlations among variables while maintaining computationally tractable inference, or they do not allow to explicitly model known correlations. To address this shortcoming, we introduce `Graph Structured Prediction Energy Networks,' for which we develop inference techniques that allow to both model explicit local and implicit higher-order correlations while maintaining tractability of inference. We apply the proposed method to tasks from the natural language processing and computer vision domain and demonstrate its general utility.