期刊
IEEE TRANSACTIONS ON ELECTRON DEVICES
卷 68, 期 10, 页码 4925-4932出版社
IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC
DOI: 10.1109/TED.2021.3108479
关键词
Resistance; Deep learning; Switches; Integrated circuit modeling; Task analysis; Noise measurement; Hafnium oxide; Binarized neural networks (BNNs); deep learning; in-memory computing; resistive random access memory (RRAM); weak reset
资金
- European Research Council Starting Grant NANOINFER [715872]
- Agence Nationale de la Recherche Grant NEURONIC [ANR-18-CE24-0009]
This study presents a model of weak RESET process in hafnium oxide RRAM for deep learning and validates its effectiveness through experiments with hybrid CMOS/RRAM technology. The model is used to train binarized neural networks for image recognition tasks, showing that device-to-device variability is the most detrimental imperfection affecting the training process.
The implementation of current deep learning training algorithms is power-hungry, due to data transfer between memory and logic units. Oxide-based resistive random access memories (RRAMs) are outstanding candidates to implement in-memory computing, which is less power-intensive. Their weak RESET regime is particularly attractive for learning, as it allows tuning the resistance of the devices with remarkable endurance. However, the resistive change behavior in this regime suffers from many fluctuations and is particularly challenging to model, especially in a way compatible with tools used for simulating deep learning. In this work, we present a model of the weak RESET process in hafnium oxide RRAM and integrate this model within the PyTorch deep learning framework. Validated on experiments on a hybrid CMOS/RRAM technology, our model reproduces both the noisy progressive behavior and the device-to-device (D2D) variability. We use this tool to train binarized neural networks (BNNs) for the MNIST handwritten digit recognition task and the CIFAR-10 object classification task. We simulate our model with and without various aspects of device imperfections to understand their impact on the training process and identify that the D2D variability is the most detrimental aspect. The framework can be used in the same manner for other types of memories to identify the device imperfections that cause the most degradation, which can, in turn, be used to optimize the devices to reduce the impact of these imperfections.
作者
我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。
推荐
暂无数据