[HTML][HTML] Analog architectures for neural network acceleration based on non-volatile memory

TP Xiao, CH Bennett, B Feinberg, S Agarwal… - Applied Physics …, 2020 - pubs.aip.org
Analog hardware accelerators, which perform computation within a dense memory array,
have the potential to overcome the major bottlenecks faced by digital hardware for data …

Improving the robustness of analog deep neural networks through a Bayes-optimized noise injection approach

N Ye, L Cao, L Yang, Z Zhang, Z Fang, Q Gu… - Communications …, 2023 - nature.com
Analog deep neural networks (DNNs) provide a promising solution, especially for
deployment on resource-limited platforms, for example in mobile settings. However, the …

An accurate, error-tolerant, and energy-efficient neural network inference engine based on SONOS analog memory

TP Xiao, B Feinberg, CH Bennett… - … on Circuits and …, 2022 - ieeexplore.ieee.org
We demonstrate SONOS (silicon-oxide-nitride-oxide-silicon) analog memory arrays that are
optimized for neural network inference. The devices are fabricated in a 40nm process and …

Benchmarking inference performance of deep learning models on analog devices

O Fagbohungbe, L Qian - 2021 International Joint Conference …, 2021 - ieeexplore.ieee.org
Analog hardware implemented deep learning models are promising for computation and
energy constrained systems such as edge computing devices. However, the analog nature …

In-memory computing array using 40nm multibit SONOS achieving 100 TOPS/W energy efficiency for deep neural network edge inference accelerators

V Agrawal, V Prabhakar, K Ramkumar… - 2020 IEEE …, 2020 - ieeexplore.ieee.org
40nm SONOS (Si-Oxide-Nitride-Oxide-Si) based non-volatile memory (NVM) cell has been
evaluated for analog memory to perform in-memory neuromorphic computing. Process flow …

The Effect of Batch Normalization on Noise Resistant Property of Deep Learning Models

O Fagbohungbe, L Qian - IEEE Access, 2022 - ieeexplore.ieee.org
The fast execution speed and energy efficiency of analog hardware have made them a
strong contender for deploying deep learning models at the edge. However, there are …

Ionizing radiation effects in SONOS-based neuromorphic inference accelerators

TP Xiao, CH Bennett, S Agarwal… - … on Nuclear Science, 2021 - ieeexplore.ieee.org
We evaluate the sensitivity of neuromorphic inference accelerators based on silicon-oxide-
nitride-oxide-silicon (SONOS) charge trap memory arrays to total ionizing dose (TID) effects …

Effect of layer-specific synaptic retention characteristics on the accuracy of deep neural networks

HN Yoo, MK Park, BG Park, JH Lee - Solid-State Electronics, 2023 - Elsevier
By modeling the change of weight/bias over time due to the retention behavior of charge trap
device (CTD), we study the influence of synaptic retention characteristics on the inference …

Impact of TID on the Analog Conductance and Training Accuracy of CBRAM-Based Neural Accelerator

P Apsangi, N Chamele, HJ Barnaby… - … on Nuclear Science, 2023 - ieeexplore.ieee.org
The changes caused by total ionizing dose (TID) in the conductance of the analog response
of Ag-Ge30Se70 conductive bridge random access memory (CBRAM) based synapses are …

A binary-activation, multi-level weight RNN and training algorithm for ADC-/DAC-free and noise-resilient processing-in-memory inference with eNVM

S Ma, D Brooks, GY Wei - IEEE Transactions on Emerging …, 2023 - ieeexplore.ieee.org
We propose a new algorithm for training neural networks with binary activations and multi-
level weights, which enables efficient processing-in-memory circuits with embedded …