Exploring Active Machine Learning Techniques to Boost Classification Accuracy in Image and Text Models

This study provides a comprehensive analysis of active machine learning algorithms designed to improve the classification precision of image and text models. This study introduces the ALiPy (Active Learning in Python) library, which provides a platform for readily integrating active learning algorithms into classification models. The proposed method enhances model performance and reduces annotation costs by selecting informative cases for annotation using the query-by-committee (QBC) procedure. The technique’s efficacy is evaluated utilizing the AG News text dataset and the CIFAR-10 image dataset. Experiment results indicate that using active learning strategies significantly enhances classification accuracy in both domains. The ALiPy library can be used by academics and professionals to research and evaluate active learning methodologies. This research advances the implementation of active learning techniques and demonstrates how this can improve the classification accuracy of image and text models. The findings demonstrate the importance of employing active learning strategies to maximize data annotation efforts and enhance the functionality of machine learning models in real-world settings.