General Transform: A Unified Framework for Adaptive Transform to Enhance Representations
This addresses the challenge of transform selection in machine learning when dataset properties are unknown, though it appears incremental as an adaptive extension of existing transform methods.
The paper tackles the problem of selecting appropriate discrete transforms for machine learning by proposing General Transform (GT), an adaptive transform that learns data-driven mappings, and demonstrates that models with GT outperform conventional transform-based approaches across computer vision and natural language processing tasks.
Discrete transforms, such as the discrete Fourier transform, are widely used in machine learning to improve model performance by extracting meaningful features. However, with numerous transforms available, selecting an appropriate one often depends on understanding the dataset's properties, making the approach less effective when such knowledge is unavailable. In this work, we propose General Transform (GT), an adaptive transform-based representation designed for machine learning applications. Unlike conventional transforms, GT learns data-driven mapping tailored to the dataset and task of interest. Here, we demonstrate that models incorporating GT outperform conventional transform-based approaches across computer vision and natural language processing tasks, highlighting its effectiveness in diverse learning scenarios.