Resource constrained neural network training
M Pietrołaj, M Blok - Scientific Reports, 2024 - nature.com
Modern applications of neural-network-based AI solutions tend to move from datacenter
backends to low-power edge devices. Environmental, computational, and power constraints …
backends to low-power edge devices. Environmental, computational, and power constraints …
[HTML][HTML] Recent implications towards sustainable and energy efficient AI and big data implementations in cloud-fog systems: A newsworthy inquiry
Cloud-fog based industries are entailing today greedy energy costs, given the wide
multiplication of their AI models and distributed BD frameworks implementations. This paper …
multiplication of their AI models and distributed BD frameworks implementations. This paper …
Self-attention mechanism to enhance the generalizability of data-driven time-series prediction: A case study of intra-hour power forecasting of urban distributed …
The emergence of small-scale urban distributed solar generation (DSG) has urged the
exploration of site-adaptive forecasting models designed to accurately predict future power …
exploration of site-adaptive forecasting models designed to accurately predict future power …
Global Stability of Phase-Change Neural Networks With Mixed Time Delays
T Dong, Y Song, H Li, X Wang… - IEEE Transactions on …, 2024 - ieeexplore.ieee.org
Phase-change memory (PCM) is a novel type of nonvolatile memory and is suitable for
artificial neural synapses. This article investigates the Lagrange global exponential stability …
artificial neural synapses. This article investigates the Lagrange global exponential stability …
Physical synthesis for advanced neural network processors
The remarkable breakthroughs in deep learning have led to a dramatic thirst for
computational resources to tackle interesting real-world problems. Various neural network …
computational resources to tackle interesting real-world problems. Various neural network …
Effective stack wear leveling for nvm
With the rapid growth of data processed by computer systems, nonvolatile memory (NVM),
represented by phase change memory (PCM), is regarded as a promising next-generation …
represented by phase change memory (PCM), is regarded as a promising next-generation …
[PDF][PDF] Review of Optimal Convolutional Neural Network Accelerator Platforms for Mobile Devices.
H Kim - J. Comput. Sci. Eng., 2022 - jcse.kiise.org
In recent years, convolutional neural networks (CNNs) have achieved remarkable
performance enhancement, and researchers have endeavored to use CNN applications on …
performance enhancement, and researchers have endeavored to use CNN applications on …
Synapse: Synergizing Approximate STT-MRAM and CNN Features for Energy-Efficient Accelerators
P Toutounchian, S Hessabi - IEEE Transactions on Sustainable …, 2025 - ieeexplore.ieee.org
Convolutional Neural Networks (CNNs) require a lot of data and parameters for accuracy.
Therefore, CNN accelerators need an efficient and extensive on-chip memory. Nevertheless …
Therefore, CNN accelerators need an efficient and extensive on-chip memory. Nevertheless …
Bit-width reduction in write counters for wear leveling in a phase-change memory system
Phase-change memory (PCM) has garnered attention as a next-generation memory owing
to its non-volatility and scalability. However, PCM wears out under excessive write accesses; …
to its non-volatility and scalability. However, PCM wears out under excessive write accesses; …
Mixture of deterministic and stochastic quantization schemes for lightweight CNN
S Kim, H Kim - 2020 International SoC Design Conference …, 2020 - ieeexplore.ieee.org
There has been a breakthrough in the field of image classification and object detection,
owing to the development of GPU and deep learning. However, because of the huge …
owing to the development of GPU and deep learning. However, because of the huge …