An Energy-Efficient Bayesian Neural Network Implementation Using Stochastic Computing Method.

The robustness of Bayesian neural networks (BNNs) to real-world uncertainties and incompleteness has led to their application in some safety-critical fields. However, evaluating uncertainty during BNN inference requires repeated sampling and feed-forward computing, making them challenging to deploy in low-power or embedded devices. This article proposes the use of stochastic computing (SC) to optimize the hardware performance of BNN inference in terms of energy consumption and hardware utilization. The proposed approach adopts bitstream to represent Gaussian random number and applies it in the inference phase. This allows for the omission of complex transformation computations in the central limit theorem-based Gaussian random number generating (CLT-based GRNG) method and the simplification of multipliers as and operations. Furthermore, an asynchronous parallel pipeline calculation technique is proposed in computing block to enhance operation speed. Compared with conventional binary radix-based BNN, SC-based BNN (StocBNN) realized by FPGA with 128-bit bitstream consumes much less energy consumption and hardware resources with less than 0.1% accuracy decrease when dealing with MNIST/Fashion-MNIST datasets.

Comparison of prognosis between emergency and scheduled hemodialysis.

BACKGROUND: The survival rate of patients undergoing hemodialysis and other renal replacement therapies has been extensively studied, but comparative studies of emergency and scheduled hemodialysis are limited. METHODS: This study included 312 patients who underwent emergency hemodialysis and 274 who received scheduled hemodialysis. We investigated the prognostic differences between these two groups of patients, including the short-term and long-term survival rates. RESULTS: The overall survival rate was significantly better among the patients in the scheduled hemodialysis group than emergency hemodialysis group. The mortality rate within 3 months of emergency hemodialysis was 4.8%, while that within 3 months of scheduled hemodialysis was 1.1%. CONCLUSIONS: Significant differences were present between emergency and scheduled hemodialysis, especially the levels of serum creatinine and hemoglobin.

Ceramide Production Mediates Aldosterone-Induced Human Umbilical Vein Endothelial Cell (HUVEC) Damages.

Here, we studied the underlying mechanism of aldosterone (Aldo)-induced vascular endothelial cell damages by focusing on ceramide. We confirmed that Aldo (at nmol/L) inhibited human umbilical vein endothelial cells (HUVEC) survival, and induced considerable cell apoptosis. We propose that ceramide (mainly C18) production might be responsible for Aldo-mediated damages in HUVECs. Sphingosine-1-phosphate (S1P), an anti-ceramide lipid, attenuated Aldo-induced ceramide production and following HUVEC damages. On the other hand, the glucosylceramide synthase (GCS) inhibitor PDMP or the ceramide (C6) potentiated Aldo-induced HUVEC apoptosis. Eplerenone, a mineralocorticoid receptor (MR) antagonist, almost completely blocked Aldo-induced C18 ceramide production and HUVEC damages. Molecularly, ceramide synthase 1 (CerS-1) is required for C18 ceramide production by Aldo. Knockdown of CerS-1 by targeted-shRNA inhibited Aldo-induced C18 ceramide production, and protected HUVECs from Aldo. Reversely, CerS-1 overexpression facilitated Aldo-induced C18 ceramide production, and potentiated HUVEC damages. Together, these results suggest that C18 ceramide production mediates Aldo-mediated HUVEC damages. MR and CerS-1 could be the two signaling molecule regulating C18 ceramide production by Aldo.