Pretrained Transform-Based Tensor Representations
来源：
发布时间：2026-08-25
报告时间：- 2026年8月28日（周五）上午10:00-11:00
- 报告地点：
- 新文科楼4楼413会议室
- 报告人：
- 赵熙乐
- 报告人单位：
- 电子科技大学
- 邀请人：
- 张雄军
- 摘要：
- Recently, Fourier transform-based tensor representations have been suggested for processing and analyzing multi-dimensional data arising from real-world applications, where transform and the corresponding latent tensor serve as the two building blocks of this framework. From both perspectives, this talk will review recent progress in transform-based tensor representations. On one hand, we will discuss the evolution of the transform from linear and shallow ones to nonlinear and deep ones. On the other hand, we will discuss the evolution of the corresponding latent tensor from shallow and untrained ones to deep and pretrained ones Finally, we will also discuss the limitations and future possibilities of the transform-based tensor representation.
赵熙乐（电子科技大学）
