Skip to main content

Multiscale Adaptive Representation of Signals: I. The Basic Framework

Author(s): Tai, Cheng; E, Weinan

To refer to this page use:
Abstract: We introduce a framework for designing multi-scale, adaptive, shift-invariant frames and bi-frames for representing signals. The new framework, called AdaFrame, improves over dictionary learning-based techniques in terms of computational efficiency at inference time. It improves classical multi-scale basis such as wavelet frames in terms of coding efficiency. It provides an attractive alternative to dictionary learning-based techniques for low level signal processing tasks, such as compression and denoising, as well as high level tasks, such as feature extraction for object recognition. Connections with deep convolutional networks are also discussed. In particular, the proposed framework reveals a drawback in the commonly used approach for visualizing the activations of the intermediate layers in convolutional networks, and suggests a natural alternative.
Publication Date: Jan-2016
Electronic Publication Date: Aug-2016
Citation: Tai, Cheng, E, Weinan. (2016). Multiscale Adaptive Representation of Signals: I. The Basic Framework. JOURNAL OF MACHINE LEARNING RESEARCH, 17
ISSN: 1532-4435
Pages: 4875-4912
Type of Material: Journal Article
Version: Author's manuscript

Items in OAR@Princeton are protected by copyright, with all rights reserved, unless otherwise indicated.