Efficient Calibration for RRAM-based In-Memory Computing using DoRA

Weirong Dong, Kai Zhou, Zhen Kong, Quan Cheng, Junkai Huang, Zhengke Yang, Masanori Hashimoto, Longyang Lin · 2025

Read Efficient Calibration for RRAM-based In-Memory Computing using DoRA by Weirong Dong, Kai Zhou, Zhen Kong, Quan Cheng, Junkai Huang, Zhengke Yang, Masanori Hashimoto, Longyang Lin (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2504-03763.txt.

Description

Weirong Dong and seven co-workers wrote a 2025 paper on fixing accuracy loss in chips that compute inside memory. Resistive memory chips are very efficient for edge AI tasks. Their stored values drift over time, though, and accuracy falls. Retraining is costly in energy, speed and chip wear.

Their fix borrows a low-rank tuning method and stores a few correction values in fast static memory. The main weights stay untouched, so no costly rewrites happen in the field. On a fifty-layer vision network, the fix restored accuracy to near seventy percent. It used just ten sample images and touched a tiny share of values.

Catalog record

CREATORWeirong Dong, Kai Zhou, Zhen Kong, Quan Cheng, Junkai Huang, Zhengke Yang, Masanori Hashimoto, Longyang Lin
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE29 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2504-03763

Browse the full library · About the Tower of Babel