Fast Deep Predictive Coding Networks for Videos Feature Extraction without Labels

TL;DR

This paper introduces a fast, unsupervised deep predictive coding network for video feature extraction that achieves high sparsity and clustering accuracy, enabling label-free object recognition in videos.

Contribution

It proposes a novel DPCN with rapid inference and a rigorous learning framework, improving sparsity and clustering over previous models.

Findings

Outperforms previous DPCNs in learning rate, sparsity, and clustering accuracy.

Validated on CIFAR-10, Super Mario Bros, and Coil-100 datasets.

Enables label-free object recognition in videos.

Abstract

Brain-inspired deep predictive coding networks (DPCNs) effectively model and capture video features through a bi-directional information flow, even without labels. They are based on an overcomplete description of video scenes, and one of the bottlenecks has been the lack of effective sparsification techniques to find discriminative and robust dictionaries. FISTA has been the best alternative. This paper proposes a DPCN with a fast inference of internal model variables (states and causes) that achieves high sparsity and accuracy of feature clustering. The proposed unsupervised learning procedure, inspired by adaptive dynamic programming with a majorization-minimization framework, and its convergence are rigorously analyzed. Experiments in the data sets CIFAR-10, Super Mario Bros video game, and Coil-100 validate the approach, which outperforms previous versions of DPCNs on learning rate, sparsity ratio, and feature clustering accuracy. Because of DCPN's solid foundation and explainability, this advance opens the door for general applications in object recognition in video without labels.