Year: 2024
Type: article
Abstract: With the rapid expansion of large-scale data-driven systems, optimizing performance under dynamic and heterogeneous environments has become increasingly challenging. Traditional model-centric approaches primarily emphasize architecture design while overlooking the intrinsic properties of data distributions, leading to limited generalization and robustness. To address this issue, this paper proposes a data-centric deep learning framework that focuses on data preprocessing, feature alignment, and self-supervised optimization. By integrating distribution-aware representation learning and adaptive feedback mechanisms, the proposed framework significantly improves performance under diverse conditions. Extensive experiments demonstrate superior accuracy, robustness, and efficiency compared with baseline methods. The results highlight that data-centric design is a critical paradigm for next-generation intelligent systems.
Author: Pembroke, Alistair
Institution: University of North Texas
Language: English
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: diamond
OpenAlex ID: W7220298506
