Scholarlink

<p>The emergence of artificial intelligence technologies has catalyzed a paradigm shift in computational capability when model sizes expand rapidly to reach trillions of parameters. Although various emerging memory technologies have achieved considerable progress for in-memory and neuromorphic computing, they fail to fully meet the massive demands of large-scale models. Owing to the advantage of the continuous tunability of ferroelectric domain patterns modulating synaptic weights during biological learning processes, ferroelectric materials have attracted extensive attention for constructing artificial synaptic devices. Currently, there appear four fundamental types of ferroelectric-based devices: ferroelectric capacitor memory, ferroelectric field-effect transistor, ferroelectric tunnel junction, and ferroelectric domain wall memory. This paper provides an in-depth analysis of their latest research progress, application areas, respective advantages, and the challenges they face. Finally, a special focus will be given to the further optimizations of material and device performance, array architecture design, neuromorphic computing architecture, and the expansion of their novel applications which are expected to be critical focal points for future research.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Thriving in-memory computing and neuromorphic applications of ferroelectric-based devices

  • 期刊论文

  • 发表日期：

    2026年2月27日

查看全文
  • Xinglong Wang,
  • Kangping Songjian,
  • Zhongfeng Ning,
  • Wendi Zhang,
  • Anquan Jiang

摘要

The emergence of artificial intelligence technologies has catalyzed a paradigm shift in computational capability when model sizes expand rapidly to reach trillions of parameters. Although various emerging memory technologies have achieved considerable progress for in-memory and neuromorphic computing, they fail to fully meet the massive demands of large-scale models. Owing to the advantage of the continuous tunability of ferroelectric domain patterns modulating synaptic weights during biological learning processes, ferroelectric materials have attracted extensive attention for constructing artificial synaptic devices. Currently, there appear four fundamental types of ferroelectric-based devices: ferroelectric capacitor memory, ferroelectric field-effect transistor, ferroelectric tunnel junction, and ferroelectric domain wall memory. This paper provides an in-depth analysis of their latest research progress, application areas, respective advantages, and the challenges they face. Finally, a special focus will be given to the further optimizations of material and device performance, array architecture design, neuromorphic computing architecture, and the expansion of their novel applications which are expected to be critical focal points for future research.