eCommons

 

Co-Designing Model Compression Algorithms and Hardware Accelerators for Efficient Deep Learning

Other Titles

Abstract

Over the past decade, machine learning (ML) with deep neural networks (DNNs) has become extremely successful in a variety of application domains including computer vision, natural language processing, and game AI. DNNs are now a primary topic of academic research among computer scientists, and a key component of commercial technologies such as web search, recommendation systems, and self-driving vehicles. However, factors such as the growing complexity of DNN models, the diminished benefits of technology scaling, and the proliferation of resource-constrained edge devices are driving a demand for higher DNN performance and energy efficiency. Consequently, neural network training and inference have begun to shift from commodity general-purpose processors (e.g., CPUs and GPUs) to custom-built hardware accelerators (e.g., FPGAs and ASICs). In line with this trend, there has been extensive research on specialized algorithms and architectures for dedicated DNN processors. Furthermore, the rapid pace of innovation in DNN algorithm space is mismatched with the time-consuming process of hardware implementation. This has generated increased interest in novel design methodologies and tools which can reduce the human effort and turn-around time of hardware design. This thesis studies how low-precision quantization and structured matrices can improve the performance and energy efficiency of DNNs running on specialized accelerators. We co-design both the DNN compression algorithms and the accelerator architectures, enabling us to evaluate the impact of our ideas on real hardware. In the process, we examine the use of high-level synthesis tools in reducing the hardware design effort. This thesis represents a cross-domain research effort at efficient deep learning. First, we propose specialized architectures for accelerating binarized neural networks on FPGA. Second, we study novel high-level synthesis techniques to reduce the manual effort in FPGA accelerator design. Third, we show a fundamental link between group convolutions and circulant matrices, two previously disparate lines of research in DNN compression. Using this insight we propose HadaNet, an alternative to circulant compression which achieve identical accuracy with asymptotically fewer multiplications. Fourth, we present outlier channel splitting, a technique to improve DNN weight quantization by removing outliers from the weight distribution without arduous retraining. Finally, we show preliminary results on overwrite quantization, a technique which address outliers in DNN activation quantization using extremely lightweight architectural extensions to a spatial accelerator template.

Journal / Series

Volume & Issue

Description

130 pages

Sponsorship

Date Issued

2020-05

Publisher

Keywords

Computer Architecture; FPGA; Hardware accelerators; Machine Learning

Location

Effective Date

Expiration Date

Sector

Employer

Union

Union Local

NAICS

Number of Workers

Committee Chair

Zhang, Zhiru

Committee Co-Chair

Committee Member

Batten, Christopher
De Sa, Christopher
Studer, Christoph

Degree Discipline

Electrical and Computer Engineering

Degree Name

Ph. D., Electrical and Computer Engineering

Degree Level

Doctor of Philosophy

Related Version

Related DOI

Related To

Related Part

Based on Related Item

Has Other Format(s)

Part of Related Item

Related To

Related Publication(s)

Link(s) to Related Publication(s)

References

Link(s) to Reference(s)

Previously Published As

Government Document

ISBN

ISMN

ISSN

Other Identifiers

Rights

Attribution-ShareAlike 4.0 International

Types

dissertation or thesis

Accessibility Feature

Accessibility Hazard

Accessibility Summary

Link(s) to Catalog Record