Google has recently introduced a new multilingual text vectorizer called RETVec (an acronym for Resilient and Efficient Text Vectorizer), to aid identification of potentially malicious content like spam and fraudulent emails in Gmail
https://security.googleblog.com/2023/11/
RETVec is a next-gen text vectorizer designed to be efficient, multilingual, and provide built-in adversarial resilience using robust word embeddings trained with similarity learning
read the paper here
huge platforms like Gmail and YouTube rely on text classification models to spot phishing attacks, inappropriate comments, and scams, threat actors are known to devise counter-strategies to bypass these defense measures
They have been observed resorting to adversarial text manipulations, which range from the use of homoglyphs to keyword stuffing to invisible characters.
RETVec aims to help build more resilient and efficient server-side and on-device text classifiers
https://thehackernews.com/2023/11/google-unveils-retvec-gmails-new.html