Blending Learning and Inference in Structured Prediction
This addresses the efficiency bottleneck in structured prediction for computer vision applications, though it appears incremental as it builds on existing graphical model frameworks.
The paper tackles the problem of learning parameters for structured predictors in graphical models by blending learning and inference, resulting in a significant speedup over traditional methods like conditional random fields and structured support vector machines, with state-of-the-art results demonstrated in tasks such as stereo estimation and semantic segmentation.
In this paper we derive an efficient algorithm to learn the parameters of structured predictors in general graphical models. This algorithm blends the learning and inference tasks, which results in a significant speedup over traditional approaches, such as conditional random fields and structured support vector machines. For this purpose we utilize the structures of the predictors to describe a low dimensional structured prediction task which encourages local consistencies within the different structures while learning the parameters of the model. Convexity of the learning task provides the means to enforce the consistencies between the different parts. The inference-learning blending algorithm that we propose is guaranteed to converge to the optimum of the low dimensional primal and dual programs. Unlike many of the existing approaches, the inference-learning blending allows us to learn efficiently high-order graphical models, over regions of any size, and very large number of parameters. We demonstrate the effectiveness of our approach, while presenting state-of-the-art results in stereo estimation, semantic segmentation, shape reconstruction, and indoor scene understanding.