Low-Rank Adaptation of Neural Fields Journal: TOG CCS: Computing methodologies Neural networks

Anh Truong, Ahmed H. Mahmoud, Mina Konaković Luković, Justin Solomon · 2025

Read Low-Rank Adaptation of Neural Fields Journal: TOG CCS: Computing methodologies Neural networks by Anh Truong, Ahmed H. Mahmoud, Mina Konaković Luković, Justin Solomon (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-2504-15933.txt.

Description

Anh Truong and co-authors adapt a trick from large language models to neural fields in a 2025 paper. Neural fields are networks that stand in for images, shapes, or other visual and physical functions.

Small changes to such fields, like filtering an image or smoothing a surface, are common but wasteful to store in full. The team uses low rank updates, which encode a change with very few numbers and no big pre trained model. Tests covered image filtering, shape editing, video compression, and energy based editing. The updates stayed light while still capturing each change well, which shows the method works across many tasks.

Catalog record

CREATORAnh Truong, Ahmed H. Mahmoud, Mina Konaković Luković, Justin Solomon
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE38 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2504-15933

Browse the full library · About the Tower of Babel