Beyond a Single Perspective: Text Anomaly Detection with Multi-View Language Representations
Yixin Liu, Kehan Yan, Shiyuan Li, Qingfeng Chen, Shirui Pan · 2026
Read Beyond a Single Perspective: Text Anomaly Detection with Multi-View Language Representations by Yixin Liu, Kehan Yan, Shiyuan Li, Qingfeng Chen, Shirui Pan (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-2601-17786.txt.
Description
Yixin Liu, Kehan Yan, Shiyuan Li, Qingfeng Chen, and Shirui Pan work on spotting odd or harmful text. Their 2026 preprint deals with text anomaly detection. The task supports content review, phishing alerts, and spam filtering. Most top methods work in two steps. First a model turns text into a set of numbers, then a detector flags the odd ones out.
The weak point, the authors argue, is the use of a single model for that first step. One view of a text can miss signs that another view would catch. Their approach draws on several views of the text at once. It also adapts better across different data sets. The goal is steadier detection when the text type changes.
Catalog record
| CREATOR | Yixin Liu, Kehan Yan, Shiyuan Li, Qingfeng Chen, Shirui Pan |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 57 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2601-17786