Crossmark

Co-optimization of LoRA rank in iteratively error injected LoRA (Ei-LoRA) and inference accuracy for tiny & error-robust edge Vision Transformer Computation-in-Memory (ViT CiM)

Crossref DOI link: https://doi.org/10.35848/1347-4065/ae480c

Published Online: 2026-04-01

Published Print: 2026-04-17

Update policy: https://doi.org/10.1088/crossmark-policy

Authors

Misawa, Naoko https://orcid.org/0009-0009-2067-5420
Zhang, Ruiqi

Wang, Tao

Matsui, Chihiro https://orcid.org/0000-0003-4594-6839
Takeuchi, Ken https://orcid.org/0000-0002-9345-6503
License Information

Version of Record valid from 2026-04-01

Text and Data Mining valid from 2026-04-01
More Information

Article Title: Co-optimization of LoRA rank in iteratively error injected LoRA (Ei-LoRA) and inference accuracy for tiny & error-robust edge Vision Transformer Computation-in-Memory (ViT CiM)

Journal Title: Japanese Journal of Applied Physics

Article Type: paper

Copyright Information: © 2026 The Japan Society of Applied Physics. All rights, including for text and data mining, AI training, and similar technologies, are reserved.

Publication dates

Date Received: 2025-11-04

Date Accepted: 2026-02-18

Online publication date: 2026-04-01

Document is current

Any future updates will be listed below