aiqbench analysis
This research introduces MIMO, a novel framework for multilingual information retrieval that resolves the tension between cross-lingual alignment and embedding uniformity using knowledge distillation from a stable English semantic anchor. By jointly optimizing distillation and contrastive learning objectives, the method significantly outperforms existing baselines in mixed-language search environments while maintaining competitive efficiency. The work advances the state of multilingual AI models by providing a robust training strategy that prevents language clustering without sacrificing retrieval discrimination.
MIMO: Multilingual Information Retrieval via Monolingual Objectives
arXiv:2605.31171v2 Announce Type: replace-cross Abstract: Multilingual Information Retrieval (MLIR) reflects real-world search environments in which queries and relevant documents may appear in different languages within a mixed-language corpus. However, existing embedding models are primarily optimized for Multi-Monolingual retrieval and their performance often degrades in MLIR settings. Moreover, directly applying conventional contrastive learning to MLIR can exacerbate language clustering and expose a trade-off between cross-lingual alignment and embedding uniformity. To address these limitations, we propose MIMO: Multilingual Information Retrieval via Monolingual Objectives, a two-stage framework that uses a stable English semantic space from a high-performing teacher model as an anchor. MIMO first initializes the student model's cross-lingual alignment through knowledge distillation, and then jointly optimizes distillation and cross-lingual contrastive learning to improve retrieval discrimination while preserving alignment. Extensive experiments show that MIMO consistently outperforms existing cross-lingual training baselines across various MLIR and Multi-Monolingual benchmarks. MIMO also remains competitive with off-the-shelf models of similar or larger parameter scales. Furthermore, our cross-lingual Alignment-Uniformity analysis clarifies the distinct roles of the two loss components and shows that their combination yields a favorable trade-off between alignment and uniformity.
Abstract:Multilingual Information Retrieval (MLIR) reflects real-world search environments in which queries and relevant documents may appear in different languages within a mixed-language corpus. However, existing embedding models are primarily optimized for Multi-Monolingual retrieval and their performance often degrades in MLIR settings. Moreover, directly applying conventional contrastive learning to MLIR can exacerbate language clustering and expose a trade-off between cross-lingual alignment and embedding uniformity. To address these limitations, we propose MIMO: Multilingual Information Retrieval via Monolingual Objectives, a two-stage framework that uses a stable English semantic space from a high-performing teacher model as an anchor. MIMO first initializes the student model's cross-lingual alignment through knowledge distillation, and then jointly optimizes distillation and cross-lingual contrastive learning to improve retrieval discrimination while preserving alignment. Extensive experiments show that MIMO consistently outperforms existing cross-lingual training baselines across various MLIR and Multi-Monolingual benchmarks. MIMO also remains competitive with off-the-shelf models of similar or larger parameter scales. Furthermore, our cross-lingual Alignment-Uniformity analysis clarifies the distinct roles of the two loss components and shows that their combination yields a favorable trade-off between alignment and uniformity.
Information Retrieval (cs.IR); Artificial Intelligence (cs.AI)
arXiv:2605.31171 [cs.IR]
(or arXiv:2605.31171v2 [cs.IR] for this version)
https://doi.org/10.48550/arXiv.2605.31171 Focus to learn more arXiv-issued DOI via DataCite
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Full text under the publisher's open licence · 273 words
Related coverage
ChemCLIR-Bench: Benchmarking Cross-Lingual Information Retrieval in Multilingual Chemical Patents
arXiv · Sep 22, 2026
[AINews] Agents disprove Polysemanticity; greatly improve Interpretability & Alignment
Latent Space · Oct 16, 2026
Disentangled Contrastive Learning for Zero-Shot Multilingual Dense Retrieval
arXiv · Aug 4, 2026
Introducing Mistral Large 4
Mistral AI · Oct 6, 2026