Deep Dependency Networks and Advanced Inference Schemes for Multi-Label Classification

TL;DR

This paper introduces deep dependency networks (DDNs), a unified framework combining dependency networks with deep learning for multi-label classification, and proposes novel inference schemes to improve label prediction accuracy on image and video datasets.

Contribution

The paper presents a new deep dependency network framework and innovative inference methods, enhancing multi-label classification performance over existing neural and Markov network approaches.

Findings

DDNs outperform basic neural architectures.

Proposed inference schemes improve label assignment accuracy.

Experimental results on multiple datasets validate the approach.

Abstract

We present a unified framework called deep dependency networks (DDNs) that combines dependency networks and deep learning architectures for multi-label classification, with a particular emphasis on image and video data. The primary advantage of dependency networks is their ease of training, in contrast to other probabilistic graphical models like Markov networks. In particular, when combined with deep learning architectures, they provide an intuitive, easy-to-use loss function for multi-label classification. A drawback of DDNs compared to Markov networks is their lack of advanced inference schemes, necessitating the use of Gibbs sampling. To address this challenge, we propose novel inference schemes based on local search and integer linear programming for computing the most likely assignment to the labels given observations. We evaluate our novel methods on three video datasets (Charades, TACoS, Wetlab) and three image datasets (MS-COCO, PASCAL VOC, NUS-WIDE), comparing their performance with (a) basic neural architectures and (b) neural architectures combined with Markov networks equipped with advanced inference and learning techniques. Our results demonstrate the superiority of our new DDN methods over the two competing approaches.