Hardware approximate techniques for deep neural network accelerators: A survey

G Armeniakos, G Zervakis, D Soudris… - ACM Computing …, 2022 - dl.acm.org
Deep Neural Networks (DNNs) are very popular because of their high performance in
various cognitive tasks in Machine Learning (ML). Recent advancements in DNNs have …

Approximate computing: Concepts, architectures, challenges, applications, and future directions

AM Dalloo, AJ Humaidi, AK Al Mhdawi… - IEEE …, 2024 - ieeexplore.ieee.org
The unprecedented progress in computational technologies led to a substantial proliferation
of artificial intelligence applications, notably in the era of big data and IoT devices. In the …

Improving the accuracy and hardware efficiency of neural networks using approximate multipliers

MS Ansari, V Mrazek, BF Cockburn… - … Transactions on Very …, 2019 - ieeexplore.ieee.org
Improving the accuracy of a neural network (NN) usually requires using larger hardware that
consumes more energy. However, the error tolerance of NNs and their applications allow …

ALWANN: Automatic layer-wise approximation of deep neural network accelerators without retraining

V Mrazek, Z Vasícek, L Sekanina… - 2019 IEEE/ACM …, 2019 - ieeexplore.ieee.org
The state-of-the-art approaches employ approximate computing to reduce the energy
consumption of DNN hardware. Approximate DNNs then require extensive retraining …

Weight-oriented approximation for energy-efficient neural network inference accelerators

ZG Tasoulas, G Zervakis… - … on Circuits and …, 2020 - ieeexplore.ieee.org
Current research in the area of Neural Networks (NN) has resulted in performance
advancements for a variety of complex problems. Especially, embedded system applications …

An improved logarithmic multiplier for energy-efficient neural computing

MS Ansari, BF Cockburn, J Han - IEEE Transactions on …, 2020 - ieeexplore.ieee.org
Multiplication is the most resource-hungry operation in neural networks (NNs). Logarithmic
multipliers (LMs) simplify multiplication to shift and addition operations and thus reduce the …

Libraries of approximate circuits: Automated design and application in CNN accelerators

V Mrazek, L Sekanina, Z Vasicek - IEEE Journal on Emerging …, 2020 - ieeexplore.ieee.org
Libraries of approximate circuits are composed of fully characterized digital circuits that can
be used as building blocks of energy-efficient implementations of hardware accelerators …

A hardware-efficient logarithmic multiplier with improved accuracy

MS Ansari, BF Cockburn, J Han - 2019 Design, Automation & …, 2019 - ieeexplore.ieee.org
Logarithmic multipliers take the base-2 logarithm of the operands and perform multiplication
by only using shift and addition operations. Since computing the logarithm is often an …

Approximate computing for ML: State-of-the-art, challenges and visions

G Zervakis, H Saadat, H Amrouch… - Proceedings of the 26th …, 2021 - dl.acm.org
In this paper, we present our state-of-the-art approximate techniques that cover the main
pillars of approximate computing research. Our analysis considers both static and …

Energy efficient edge computing enabled by satisfaction games and approximate computing

N Irtija, I Anagnostopoulos, G Zervakis… - IEEE Transactions …, 2021 - ieeexplore.ieee.org
In this paper, we introduce an energy efficient edge computing solution to collaboratively
utilize Multi-access Edge Computing (MEC) and Fully Autonomous Aerial Systems (FAAS) to …