Optimization Approaches in Meta-Learning Models
Nidhi R. Arora, Ashok Kumar Sharma, Dinesh Kumar · Advances in computational intelligence and robotics book series · 2023
This book chapter provides a comprehensive overview of optimization approaches in meta-learning, focusing on techniques and their applications. Meta-learning is a subfield of machine learning that emphasizes acquiring knowledge from previous tasks and applying the same to new tasks in order to develop the models with improved learning process. Optimization plays a crucial role in meta-learning models by enabling the effective acquisition and utilization of knowledge across tasks. This chapter provides an overview of various optimization approaches employed in meta-learning models which entail changing the model's input parameters or learning algorithms to facilitate effective learning across various tasks or domains. The methods tackle the problem of effective learning without compromising with accuracy and precision in performance focusing on the benefits of meta-learning frameworks in practical situations which may be considered as the real-world applications of these approaches.