Accurate yet efficient stochastic computing neural acceleration with high precision residual fusion

Y Hu, T Zhang, R Wei, M Li, R Wang… - … , Automation & Test …, 2023 - ieeexplore.ieee.org
Stochastic computing (SC) emerges as a fault-tolerant and area-efficient computing
paradigm for neural acceleration. However, existing SC accelerators suffer from an intrinsic …

Efficient yet Accurate End-to-End SC Accelerator Design

M Li, Y Hu, T Zhang, R Wei, Y Zhang, R Huang… - arxiv preprint arxiv …, 2024 - arxiv.org
Providing end-to-end stochastic computing (SC) neural network acceleration for state-of-the-
art (SOTA) models has become an increasingly challenging task, requiring the pursuit of …

Efficient Non-Linear Adder for Stochastic Computing with Approximate Spatial-Temporal Sorting Network

Y Hu, T Zhang, M Li, R Wei, L Lai… - 2023 60th ACM/IEEE …, 2023 - ieeexplore.ieee.org
End-to-end stochastic computing (SC) enables fault-tolerant and area-efficient neural
acceleration by conducting non-linear addition, including accumulation and activation …

Highly accurate and energy efficient binary-stochastic multipliers for fault-tolerant applications

Y Zhang, L **e, J Han, X Cheng… - IEEE Transactions on …, 2022 - ieeexplore.ieee.org
Stochastic circuits use randomly distributed bitstreams to represent numbers, so leading to
small areas and low power dissipation. However, it does not only result in a long latency and …

A Study of Signed-Digit Hybrid Stochastic Number for Arithmetic Computing

Y Song, H Li, X Zhu, Y Chen - IEEE Transactions on Very Large …, 2025 - ieeexplore.ieee.org
In this article, a signed-digit hybrid stochastic number (SD-HSN), which combines the two-
line bipolar stochastic number (TLB-SN) and the binary signed-digit (BSD) number, is …

Not your father's stochastic computing (SC)! Efficient yet Accurate End-to-End SC Accelerator Design

M Li, Y Hu, T Zhang, R Wei, Y Zhang… - 2023 IEEE 15th …, 2023 - ieeexplore.ieee.org
Providing end-to-end stochastic computing (SC) neural network acceleration for state-of-the-
art (SOTA) models has become an increasingly challenging task, requiring the pursuit of …