Search for: All records

In-memory-computing (IMC) SRAM architecture has gained significant attention as it achieves high energy efficiency for computing a convolutional neural network (CNN) model [1]. Recent works investigated the use of analog-mixed-signal (AMS) hardware for high area and energy efficiency [2], [3]. However, AMS hardware output is well known to be susceptible to process, voltage, and temperature (PVT) variations, limiting the computing precision and ultimately the inference accuracy of a CNN. We reconfirmed, through the simulation of a capacitor-based IMC SRAM macro that computes a 256D binary dot product, that the AMS computing hardware has a significant root-mean-square error (RMSE) of 22.5% across the worst-case voltage, temperature (Fig. 16.1.1 top left) and 3-sigma process variations (Fig. 16.1.1 top right). On the other hand, we can implement an IMC SRAM macro using robust digital logic [4], which can virtually eliminate the variability issue (Fig. 16.1.1 top). However, digital circuits require more devices than AMS counterparts (e.g., 28 transistors for a mirror full adder [FA]). As a result, a recent digital IMC SRAM shows a lower area efficiency of 6368F2/b (22nm, 4b/4b weight/activation) [5] than the AMS counterpart (1170F2/b, 65nm, 1b/1b) [3]. In light of this, we aim to adopt approximate arithmetic hardware to improve area and power efficiency and present two digital IMC macros (DIMC) with different levels of approximation (Fig. 16.1.1 bottom left). Also, we propose an approximation-aware training algorithm and a number format to minimize inference accuracy degradation induced by approximate hardware (Fig. 16.1.1 bottom right). We prototyped a 28nm test chip: for a 1b/1b CNN model for CIFAR-10 and across 0.5-to-1.1V supply, the DIMC with double-approximate hardware (DIMC-D) achieves 2569F2/b, 932-2219TOPS/W, 475-20032GOPS, and 86.96% accuracy, while for a 4b/1b CNN model, the DIMC with the single-approximate hardware (DIMC-S) achieves 3814F2/b, 458-990TOPS/W 

Summary Group testing involves pooling individual specimens (e.g., blood, urine, swabs, etc.) and testing the pools for the presence of disease. When the proportion of diseased individuals is small, group testing can greatly reduce the number of tests needed to screen a population. Statistical research in group testing has traditionally focused on applications for a single disease. However, blood service organizations and large-scale disease surveillance programs are increasingly moving towards the use of multiplex assays, which measure multiple disease biomarkers at once. Tebbs and others (2013, Two-stage hierarchical group testing for multiple infections with application to the Infertility Prevention Project. Biometrics69, 1064–1073) and Hou and others (2017, Hierarchical group testing for multiple infections. Biometrics73, 656–665) were the first to examine hierarchical group testing case identification procedures for multiple diseases. In this article, we propose new non-hierarchical procedures which utilize two-dimensional arrays. We derive closed-form expressions for the expected number of tests per individual and classification accuracy probabilities and show that array testing can be more efficient than hierarchical procedures when screening individuals for multiple diseases at once. We illustrate the potential of using array testing in the detection of chlamydia and gonorrhea for a statewide screening program in Iowa. Finally, we describe an R/Shiny application that will help practitioners identify the best multiple-disease case identification algorithm.