Efficient Temporal-aware Matryoshka Adaptation for Temporal Information Retrieval
Tuan-Luc Huynh Weiqing Wang Trung Le, Thuy-Trang Vu Dragan Gašević Yuan-Fang Li Thanh-Toan Do · 2026
Read Efficient Temporal-aware Matryoshka Adaptation for Temporal Information Retrieval by Tuan-Luc Huynh Weiqing Wang Trung Le, Thuy-Trang Vu Dragan Gašević Yuan-Fang Li Thanh-Toan Do (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-2601-05549.txt.
Description
Tuan-Luc Huynh and 6 co-authors present TMRL, a 2026 method that teaches retrievers to respect time in search results. Retrieval-augmented systems fail when they fetch context from the wrong period, no matter how well the generator reasons afterward. Standard embeddings capture meaning but often blur dates and temporal order.
TMRL builds on Matryoshka embeddings, whose nested structure stores information at several sizes. It carves out a temporal subspace inside that structure, adding time awareness while keeping general meaning intact. The method adapts several standard encoder models efficiently. Tests reported temporal retrieval and temporal question-answering performance competitive with specialized approaches, while letting users trade accuracy against speed by choosing embedding size.
Catalog record
| CREATOR | Tuan-Luc Huynh Weiqing Wang Trung Le, Thuy-Trang Vu Dragan Gašević Yuan-Fang Li Thanh-Toan Do |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 60 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2601-05549