Top-Down Learning for Structured Labeling with Convolutional Pseudoprior
This addresses the problem of improving structured labeling in computer vision and NLP by offering a more automated and powerful alternative to existing methods like CRFs and RNNs, though it appears incremental as it builds on prior CNN and pseudo-likelihood techniques.
The paper tackles the unclear role of top-down processes in CNNs for structured labeling by proposing a convolutional pseudo-prior method that automatically learns contextual kernels, achieving state-of-the-art results on sequential and image labeling benchmarks.
Current practice in convolutional neural networks (CNN) remains largely bottom-up and the role of top-down process in CNN for pattern analysis and visual inference is not very clear. In this paper, we propose a new method for structured labeling by developing convolutional pseudo-prior (ConvPP) on the ground-truth labels. Our method has several interesting properties: (1) compared with classical machine learning algorithms like CRFs and Structural SVM, ConvPP automatically learns rich convolutional kernels to capture both short- and long- range contexts; (2) compared with cascade classifiers like Auto-Context, ConvPP avoids the iterative steps of learning a series of discriminative classifiers and automatically learns contextual configurations; (3) compared with recent efforts combing CNN models with CRFs and RNNs, ConvPP learns convolution in the labeling space with much improved modeling capability and less manual specification; (4) compared with Bayesian models like MRFs, ConvPP capitalizes on the rich representation power of convolution by automatically learning priors built on convolutional filters. We accomplish our task using pseudo-likelihood approximation to the prior under a novel fixed-point network structure that facilitates an end-to-end learning process. We show state-of-the-art results on sequential labeling and image labeling benchmarks.