CLeAN: C ontinual Le arning A daptive N ormalization in Dynamic Environments

Isabella Marasco, Davide Evangelista, Elena Loli Piccolomini, Michele Colajanni · 2026

Read CLeAN: C ontinual Le arning A daptive N ormalization in Dynamic Environments by Isabella Marasco, Davide Evangelista, Elena Loli Piccolomini, Michele Colajanni (2026) 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-2603-17548.txt.

Description

Marasco, Evangelista, Piccolomini, and Colajanni present CLeAN, an adaptive normalization method for continual learning on tabular data. Standard scaling needs the whole dataset at once, which streaming tasks never offer. CLeAN learns global feature scales and updates them with an exponential moving average. Tests span two datasets and replay, A-GEM, and EWC strategies. Scores on new data rose while forgetting dropped, they report. Normalization itself guards old knowledge, they argue.

Catalog record

CREATORIsabella Marasco, Davide Evangelista, Elena Loli Piccolomini, Michele Colajanni
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE59135
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2603.17548

Browse the full library · About the Tower of Babel