{"title":"MEFET-Based CAM/TCAM for Memory-Augmented Neural Networks","authors":"Sai Sanjeet;Jonathan Bird;Bibhu Datta Sahoo","doi":"10.1109/JXCDC.2024.3410681","DOIUrl":null,"url":null,"abstract":"Memory-augmented neural networks (MANNs) require large external memories to enable long-term memory storage and retrieval. Content-addressable memory (CAM) is a type of memory used for high-speed searching applications and is well-suited for MANNs. Recent advances in exploratory nonvolatile devices have spurred the development of nonvolatile CAMs. However, these devices suffer from poor ON-OFF ratio, large write voltages, and long write times. This work proposes a nonvolatile ternary CAM (TCAM) using magnetoelectric field effect transistors (MEFETs). The energy and delay of various operations are simulated using the ASAP 7-nm predictive technology for the transistors and a Verilog-A model of the MEFET. The proposed structure achieves orders of magnitude improvement in search energy and \n<inline-formula> <tex-math>$\\gt 45\\times $ </tex-math></inline-formula>\n improvement in search energy-delay product compared with prior works. The write energy and delay are also improved by \n<inline-formula> <tex-math>$8\\times $ </tex-math></inline-formula>\n and \n<inline-formula> <tex-math>$12\\times $ </tex-math></inline-formula>\n, respectively, compared with CAMs designed with other nonvolatile devices. A variability analysis is performed to study the effect of process variations on the CAM. The proposed CAM is then used to build a one-shot learning MANN and is benchmarked with the Modified National Institute of Standards and Technology (MNIST), extended MNIST (EMNIST), and labeled faces in the wild (LFW) datasets with binary embeddings, giving >99% accuracy on MNIST, a top-3 accuracy of 97.11% on the EMNIST dataset, and >97% accuracy on the LFW dataset, with embedding sizes of 16, 64, and 512, respectively. The proposed CAM is shown to be fast, energy-efficient, and scalable, making it suitable for MANNs.","PeriodicalId":54149,"journal":{"name":"IEEE Journal on Exploratory Solid-State Computational Devices and Circuits","volume":"10 ","pages":"31-39"},"PeriodicalIF":2.0000,"publicationDate":"2024-06-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=10550938","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"IEEE Journal on Exploratory Solid-State Computational Devices and Circuits","FirstCategoryId":"1085","ListUrlMain":"https://ieeexplore.ieee.org/document/10550938/","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q3","JCRName":"COMPUTER SCIENCE, HARDWARE & ARCHITECTURE","Score":null,"Total":0}
引用次数: 0
Abstract
Memory-augmented neural networks (MANNs) require large external memories to enable long-term memory storage and retrieval. Content-addressable memory (CAM) is a type of memory used for high-speed searching applications and is well-suited for MANNs. Recent advances in exploratory nonvolatile devices have spurred the development of nonvolatile CAMs. However, these devices suffer from poor ON-OFF ratio, large write voltages, and long write times. This work proposes a nonvolatile ternary CAM (TCAM) using magnetoelectric field effect transistors (MEFETs). The energy and delay of various operations are simulated using the ASAP 7-nm predictive technology for the transistors and a Verilog-A model of the MEFET. The proposed structure achieves orders of magnitude improvement in search energy and
$\gt 45\times $
improvement in search energy-delay product compared with prior works. The write energy and delay are also improved by
$8\times $
and
$12\times $
, respectively, compared with CAMs designed with other nonvolatile devices. A variability analysis is performed to study the effect of process variations on the CAM. The proposed CAM is then used to build a one-shot learning MANN and is benchmarked with the Modified National Institute of Standards and Technology (MNIST), extended MNIST (EMNIST), and labeled faces in the wild (LFW) datasets with binary embeddings, giving >99% accuracy on MNIST, a top-3 accuracy of 97.11% on the EMNIST dataset, and >97% accuracy on the LFW dataset, with embedding sizes of 16, 64, and 512, respectively. The proposed CAM is shown to be fast, energy-efficient, and scalable, making it suitable for MANNs.