Hardware Acceleration and Approximation of CNN Computations: Case Study on an Integer Version of LeNet

AI systems have an increasing sprawling impact in many application areas. Embedded systems built on AI have strong conflictual implementation constraints, including high computation speed, low power consumption, high energy efficiency, strong robustness and low cost. Neural Networks (NNs) used by th...

Full description

Saved in:
Bibliographic Details
Published in:Electronics (Basel) 2024-07, Vol.13 (14), p.2709
Main Authors: Leveugle, Régis, Cogney, Arthur, Gah El Hilal, Ahmed Baba, Lailler, Tristan, Pieau, Maxime
Format: Article
Language:English
Subjects:
Acceleration
Accuracy
Approximation
Artificial Intelligence
Artificial neural networks
Autonomous vehicles
Case studies
Computation
Computational linguistics
Computer Science
Constraints
Critical path
Embedded systems
Energy consumption
Energy efficiency
Engineering Sciences
Fault tolerance
Field programmable gate arrays
Hardware
Integers
Language processing
Micro and nanotechnologies
Microelectronics
Natural language interfaces
Neural and Evolutionary Computing
Neural networks
Pattern recognition
Power consumption
Software
Citations: Items that this one cites
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:AI systems have an increasing sprawling impact in many application areas. Embedded systems built on AI have strong conflictual implementation constraints, including high computation speed, low power consumption, high energy efficiency, strong robustness and low cost. Neural Networks (NNs) used by these systems are intrinsically partially tolerant to computation disturbances. As a consequence, they are an interesting target for approximate computing seeking reduced resources, lower power consumption and faster computation. Also, the large number of computations required by a single inference makes hardware acceleration almost unavoidable to globally meet the design constraints. The reported study, based on an integer version of LeNet, shows the possible gains when coupling approximation and hardware acceleration. The main conclusions can be leveraged when considering other types of NNs. The first one is that several approximation types that look very similar can exhibit very different trade-offs between accuracy loss and hardware optimizations, so the selected approximation has to be carefully chosen. Also, a strong approximation leading to the best hardware can also lead to the best accuracy. This is the case here when selecting the ApxFA5 adder approximation defined in the literature. Finally, combining hardware acceleration and approximate operators in a coherent manner also increases the global gains.
ISSN:2079-9292
2079-9292
DOI:10.3390/electronics13142709