Merging Embedded Topics with Optimal Transport for Online Topic Modeling on Data Streams
Federica Granese (✉), Benjamin Navet, Serena Villata, Charles Bouveyron · 2025
Read Merging Embedded Topics with Optimal Transport for Online Topic Modeling on Data Streams by Federica Granese (✉), Benjamin Navet, Serena Villata, Charles Bouveyron (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-2504-07711.txt.
Description
Federica Granese, Benjamin Navet, Serena Villata, and Charles Bouveyron wrote a 2025 paper on topic modeling for text that arrives as a stream. Social media pours out text daily, so models must update as new batches land, not retrain from scratch. Topics also drift over time, and real shifts should be flagged.
Their method, StreamETM, extends a neural topic model that embeds topics and words in one space. Models trained on successive batches are merged with an unbalanced optimal transport method, which matches topics across time at low cost. A change point detector watches the stream and marks when topics truly shift. Tests on simulated and real data showed the method beating rivals. The code is public for other researchers.
Catalog record
| CREATOR | Federica Granese (✉), Benjamin Navet, Serena Villata, Charles Bouveyron |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 58 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2504-07711