Domain Generalization Through Data Augmentation: A Survey of Methods, Applications, and Challenges

Mathematics, Journal Year: 2025, Volume and Issue: 13(5), P. 824 - 824

Published: Feb. 28, 2025

Domain generalization (DG) has become a pivotal research area in machine learning, focusing on equipping models with the ability to generalize effectively unseen test domains that differ from training distribution. This capability is crucial, as real-world data frequently exhibit domain shifts violate assumption of independent and identically distributed (i.i.d.) data, resulting significant declines model performance. Among various strategies address generalization, augmentation garnered substantial attention an effective approach for mitigating improving robustness. In this survey, we examine role offering comprehensive overview its methods, applications, challenges. We present detailed taxonomy techniques, categorized along three dimensions: scope, nature, dependency. Additionally, provide comparative analysis key highlighting their strengths limitations. Finally, explore domain-specific applications analyze effectiveness enhancing across tasks, including computer vision, NLP, speech, robotics. conclude by examining challenges—such computational cost overfitting—and outline promising directions, focus advancing cross-modal techniques developing standardized evaluation benchmarks.

Language: Английский

Neuromorphic computing-enabled generalized machine fault diagnosis with dynamic vision

Advanced Engineering Informatics, Journal Year: 2025, Volume and Issue: 65, P. 103300 - 103300

Published: April 4, 2025

Language: Английский