Stochastic nanomagnets as current digitizers for efficient probabilistic machine learning
Analog in-memory computing based on crossbar arrays offers a path to energy-efficient AI hardware, but has been limited by reliance on bulky, power-hungry analog-to-digital converters. This study introduces stochastic nanomagnets driven by spin-orbit torque as intrinsic analog-to-digital interfaces, enabling compact, fast, low-power digitization while maintaining high computational accuracy. These results suggest a promising direction for energy-efficient AI hardware accelerators and significant advances in next-generation machine-learning hardware.
Authors
- Luqiao Liu (ORCID: https://orcid.org/0000-0001-6892-8102)
- Marc A. Baldo (ORCID: https://orcid.org/0000-0003-2201-5257)
- Jungsoo Lee (ORCID: https://orcid.org/0000-0001-9439-0166)
- Qiuyuan Wang (ORCID: https://orcid.org/0000-0002-4764-3383)
- Brooke C. McGoldrick
- Dooyong Koh (ORCID: https://orcid.org/0000-0002-5368-2856)
- Kshemal K. Gupte
- Cheng Wang (ORCID: https://orcid.org/0009-0008-0728-1622)
- Ethan G. Rogers
Institutions
- Iowa State University (US)
- Massachusetts Institute of Technology (US)
Publication Details
- Journal
- Physical Review Applied
- Published
- 2025-10-03
- DOI
- https://doi.org/10.1103/s86j-5jcr
- Primary Topic
- Advanced Memory and Neural Computing
- Type
- article
- Field-Weighted Citation Impact
- 0.00