Image Classification by Reinforcement Learning with Two-State Q-Learning

TL;DR

This paper introduces a simple hybrid classifier combining deep learning and two-state Q-learning for image classification, achieving superior performance on multiple datasets with fewer parameters.

Contribution

It proposes a novel two-state Q-learning method for image classification that simplifies the model and improves performance compared to existing feature map-based approaches.

Findings

Outperforms ResNet50 and InceptionV3 on ImageNet, Cats and Dogs, Caltech-101 datasets.

Uses fewer optimization parameters due to two-state Q-learning.

Achieves higher accuracy with a simpler reward function.

Abstract

In this paper, a simple and efficient Hybrid Classifier is presented which is based on deep learning and reinforcement learning. Here, Q-Learning has been used with two states and 'two or three' actions. Other techniques found in the literature use feature map extracted from Convolutional Neural Networks and use these in the Q-states along with past history. This leads to technical difficulties in these approaches because the number of states is high due to large dimensions of the feature map. Because the proposed technique uses only two Q-states it is straightforward and consequently has much lesser number of optimization parameters, and thus also has a simple reward function. Also, the proposed technique uses novel actions for processing images as compared to other techniques found in literature. The performance of the proposed technique is compared with other recent algorithms like ResNet50, InceptionV3, etc. on popular databases including ImageNet, Cats and Dogs Dataset, and Caltech-101 Dataset. The proposed approach outperforms others techniques on all the datasets used.