---
title: "When Does Mixing Help? Analyzing Query Embedding"
description: "This study asks when blending two languages in a query actually improves dense retrieval. 22 references."
authors: ["Tongyao Zhu", "Huang Chao Ming", "Min-Yen Kan"]
year: 2026
venue: "Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)"
source: "https://aclanthology.org/2026.acl-long.1455/"
doi: "10.18653/v1/2026.acl-long.1455"
licence: "CC BY 4.0 (ACL Anthology)"
licence_url: "https://creativecommons.org/licenses/by/4.0/"
summary_licence: "Elamur Cosmic License"
canonical: "https://research.elamur.ai/free/en/papers/zhu-2026-when-does-mixing-help-analyzing/"
---
[Elamur Research](https://md.research.elamur.ai/free/en/) · [this page in HTML](https://research.elamur.ai/free/en/papers/zhu-2026-when-does-mixing-help-analyzing/)

[home](https://md.research.elamur.ai/free/en/) / [papers](https://md.research.elamur.ai/free/en/papers/) / [2026](https://md.research.elamur.ai/free/en/papers/2026/page/56/) / When Does Mixing Help? Analyzing Query Embedding Interpolation in Multilingual Dense Retrieval

# When Does Mixing Help? Analyzing Query Embedding Interpolation in Multilingual Dense Retrieval

**Authors:** [Tongyao Zhu](https://md.research.elamur.ai/free/en/authors/tongyao-zhu/), Huang Chao Ming and [Min-Yen Kan](https://md.research.elamur.ai/free/en/authors/min-yen-kan/) · **Year:** 2026 · **Venue:** [Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)](https://md.research.elamur.ai/free/en/venues/acl/) ([ACL 2026](https://md.research.elamur.ai/free/en/venues/acl/2026/)) · **Publisher:** [Association for Computational Linguistics](https://md.research.elamur.ai/free/en/publishers/association-for-computational-linguistics/) · **Licence:** [CC BY 4.0 (ACL Anthology)](https://creativecommons.org/licenses/by/4.0/)

**Links:** [Publisher's page](https://aclanthology.org/2026.acl-long.1455/) · [PDF](https://aclanthology.org/2026.acl-long.1455.pdf) · [DOI 10.18653/v1/2026.acl-long.1455](https://doi.org/10.18653/v1/2026.acl-long.1455) · [OpenAlex](https://openalex.org/W7166878671)

**Citations:** 0 citations (OpenAlex, 2026-10-14)

## Summary
Mixed-language queries are common among multilingual users, yet standard retrieval benchmarks only test monolingual or fully translated queries. This study asks when blending two languages in a query actually improves dense retrieval. Producing code-mixed text with an LLM is costly and hard to control, so the authors use a simpler substitute. They encode parallel mMARCO query translations with BGE-M3 and take a weighted average (convex combination) of the two monolingual embeddings at fixed ratios. The resulting vector is re-normalized and used to search a FAISS index. They cover 35 language pairs, each evaluated against L1-only, L2-only and combined L1+L2 document collections. LLM-generated word-level mixing serves only as a validation check, and it shows similar trends. In most settings, some interior ratio beats the stronger monolingual query, but the benefit depends strongly on English. These regularities make mixing sensitivity fairly predictable, which gives concrete guidance for multilingual search and RAG systems.

## Key points
- An optimal mixing ratio outperformed the best pure-language query in 88 of 105 settings (83.8%), measured by nDCG@10 over top-100 retrieval.
- English drives an asymmetry:
  - For indices without English, mixing always helped.
  - For indices containing English, pure English queries worked best.
  - For every non-English document language, English was the most effective mixing partner.
- Setup: all 14 mMARCO languages, a filtered set of 1,484 longer queries, the full 8.8M-passage corpus for main runs, and 100k-passage subsets for ablations. Mixing weights were λ ∈ {0, 10, 30, 50, 70, 90, 100}.
- After controlling for English, gains shrink as the typological distance between the two query languages grows. Script and resource level show only weak links.
- The patterns largely hold across other model families and sizes. Code is available at github.com/tongyao-zhu/query-embedding-mix.

*Our summary of the paper, in our own words — not its text. The work belongs to its authors; read it at the links above.*

## Institutions

- [National University of Singapore](https://md.research.elamur.ai/free/en/universities/national-university-of-singapore/) — University, Singapore, Singapore · 207 works · 39,576 citations

📚 Shelves (5): [🖥️ Computer Science](https://md.research.elamur.ai/free/en/shelves/computer-science/) · 18,614 works · [🤖 Artificial Intelligence](https://md.research.elamur.ai/free/en/shelves/computer-science/artificial-intelligence/) · 17,965 works · [🎭 Arts and Humanities](https://md.research.elamur.ai/free/en/shelves/arts-and-humanities/) · 17,450 works · [🗣️ Language and Linguistics](https://md.research.elamur.ai/free/en/shelves/arts-and-humanities/language-and-linguistics/) · 17,296 works · [🗄️ Information Systems](https://md.research.elamur.ai/free/en/shelves/computer-science/information-systems/) · 1,730 works

🧭 Topics (2): [🌍 Multilingual & cross-lingual NLP](https://md.research.elamur.ai/free/en/topics/multilingual-nlp/) · 882 works · [🔍 Information retrieval & search](https://md.research.elamur.ai/free/en/topics/information-retrieval/) · 552 works

## Cite

Cite as: Tongyao Zhu, Huang Chao Ming and Min-Yen Kan. 2026. When Does Mixing Help? Analyzing Query Embedding Interpolation in Multilingual Dense Retrieval. Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers).

The paper itself, not this page. BibTeX from the library's record of the work.

```bibtex
@inproceedings{zhu2026when,
  title = {{When Does Mixing Help? Analyzing Query Embedding Interpolation in Multilingual Dense Retrieval}},
  author = {Tongyao Zhu, Huang Chao Ming and Min-Yen Kan},
  booktitle = {Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)},
  year = {2026},
  doi = {10.18653/v1/2026.acl-long.1455},
  url = {https://doi.org/10.18653/v1/2026.acl-long.1455}
}
```

- [← previous When Does Auxiliary Modality Matter in Solving Geometric Problems? A Comprehensive Study of Textual, Formal, and Visual Modalities](https://md.research.elamur.ai/free/en/papers/namgoong-2026-when-does-auxiliary-modality-matter/)
- [next → When Does OmniASR Fail? A Fine-Grained Human Evaluation on Saudi Arabic Dialects](https://md.research.elamur.ai/free/en/papers/khalifa-2026-when-does-omniasr-fail-fine/)
- [↑ back to papers of 2026](https://md.research.elamur.ai/free/en/papers/2026/page/56/)

## References

22 references, 14 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

1. Hengyi Fu. 2018. Mixed language queries in online searches: A study of intra-sentential code-switching from a qualitative perspective. Aslib Journal of Information Management, 71(1):72–89. [find](https://scholar.google.com/scholar?q=Hengyi+Fu%2E+2018%2E+Mixed+language+queries+in+online+searches%3A+A+study+of+intra%2Dsentential+code%2Dswitching+from+a+qualitative+perspective%2E+Aslib+Journal+of+Information+Management%2C+71%281%29%3A72%E2%80%9389%2E)
2. We use open-source datasets and models to study the behavior of retrieval encoders for code-mixed queries. It helps to understand the models better. We do not see any potential risk or ethical impact. [find](https://scholar.google.com/scholar?q=We+use+open%2Dsource+datasets+and+models+to+study+the+behavior+of+retrieval+encoders+for+code%2Dmixed+queries%2E+It+helps+to+understand+the+models+better%2E+We+do+not+see+any+potential+risk+or+ethical+impact%2E)
3. Gustavo Aguilar, Sudipta Kar, and Thamar Solorio. 2020. [LinCE: A centralized benchmark for linguistic code-switching evaluation](https://md.research.elamur.ai/free/en/papers/aguilar-2020-lince-centralized-benchmark-linguistic-code/). In Proceedings of the Twelfth Language Resources and Evaluation Conference, pages 1803–1813, Marseille, France. European Language Resources Association.
4. Rob Van Der Goot, Esther Ploeger, Verena Blaschke, and Tanja Samardzic. 2025. DistaLs: a comprehensive collection of language distance measures. In Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing: System Demonstrations, pages 307–318, Suzhou, China. Association… [find](https://scholar.google.com/scholar?q=Rob+Van+Der+Goot%2C+Esther+Ploeger%2C+Verena+Blaschke%2C+and+Tanja+Samardzic%2E+2025%2E+DistaLs%3A+a+comprehensive+collection+of+language+distance+measures%2E+In+Proceedings+of+the+2025+Conference+on+Empirical+Meth)
5. Gautier Izacard, Mathilde Caron, Lucas Hosseini, Sebastian Riedel, Piotr Bojanowski, Armand Joulin, and Edouard Grave. 2022. [Unsupervised dense information retrieval with contrastive learning](https://md.research.elamur.ai/free/en/papers/izacard-2021-unsupervised-dense-information-retrieval-contrastive/). Preprint, arXiv:2112.09118.
6. Pratik Joshi, Sebastin Santy, Amar Budhiraja, Kalika Bali, and Monojit Choudhury. 2020. [The state and fate of linguistic diversity and inclusion in the NLP world](https://md.research.elamur.ai/free/en/papers/joshi-2020-state-fate-linguistic-diversity-inclusion/). In Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics.
7. Vladimir Karpukhin, Barlas Oguz, Sewon Min, Patrick Lewis, Ledell Wu, Sergey Edunov, Danqi Chen, and Wen-tau Yih. 2020. [Dense passage retrieval for open-domain question answering](https://md.research.elamur.ai/free/en/papers/karpukhin-2020-dense-passage-retrieval-open-domain/). In Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing (EMNLP), pages 6769–6781. Asso…
8. Simran Khanuja, Sandipan Dandapat, Anirudh Srinivasan, Sunayana Sitaram, and Monojit Choudhury. 2020. [GLUECoS: An evaluation benchmark for code-switched NLP](https://md.research.elamur.ai/free/en/papers/khanuja-2020-gluecos-evaluation-benchmark-code-switched/). In Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics, pages 3575–3585, Online. Association for Computati…
9. Jonghwi Kim, Deokhyung Kang, Seonjeong Hwang, Yunsu Kim, Jungseul Ok, and Gary Lee. 2025. MiLQ: Benchmarking IR models for bilingual web search with mixed language queries. Preprint, arXiv:2505.16631. [find](https://scholar.google.com/scholar?q=Jonghwi+Kim%2C+Deokhyung+Kang%2C+Seonjeong+Hwang%2C+Yunsu+Kim%2C+Jungseul+Ok%2C+and+Gary+Lee%2E+2025%2E+MiLQ%3A+Benchmarking+IR+models+for+bilingual+web+search+with+mixed+language+queries%2E+Preprint%2C+arXiv%3A2505%2E16631%2E)
10. Sneha Kudugunta, Isaac Caswell, Biao Zhang, Xavier Garcia, Derrick Xin, Aditya Kusupati, Romi Stella, Ankur Bapna, and Orhan Firat. 2023. [Madlad-400: A multilingual and document-level large audited dataset](https://md.research.elamur.ai/free/en/papers/kudugunta-2023-madlad-400-multilingual-document-level/). In Advances in Neural Information Processing Systems, volume 36, pages 67284–67296. Curran As…
11. Robert Litschko, Ekaterina Artemova, and Barbara Plank. 2023. Boosting zero-shot cross-lingual retrieval by training on artificially code-switched data. In Findings of the Association for Computational Linguistics: ACL 2023, pages 3096–3108, Toronto, Canada. Association for Computational Linguistics… [find](https://scholar.google.com/scholar?q=Robert+Litschko%2C+Ekaterina+Artemova%2C+and+Barbara+Plank%2E+2023%2E+Boosting+zero%2Dshot+cross%2Dlingual+retrieval+by+training+on+artificially+code%2Dswitched+data%2E+In+Findings+of+the+Association+for+Computationa)
12. NLLB Team, Marta R. Costa-jussà, James Cross, Onur Çelebi, Maha Elbayad, Kenneth Heafield, Kevin Heffernan, Elahe Kalbassi, Janice Lam, Daniel Licht, Jean Maillard, Anna Sun, Skyler Wang, Guillaume Wenzek, Al Youngblood, Bapi Akula, Loic Barrault, Gabriel Mejia-Gonzalez, Prangthip Hansanti, and 20 o… [find](https://scholar.google.com/scholar?q=NLLB+Team%2C+Marta+R%2E+Costa%2Djuss%C3%A0%2C+James+Cross%2C+Onur+%C3%87elebi%2C+Maha+Elbayad%2C+Kenneth+Heafield%2C+Kevin+Heffernan%2C+Elahe+Kalbassi%2C+Janice+Lam%2C+Daniel+Licht%2C+Jean+Maillard%2C+Anna+Sun%2C+Skyler+Wang%2C+Guillaume+We)
13. Peng Qi, Yuhao Zhang, Yuhui Zhang, Jason Bolton, and Christopher D. Manning. 2020. [Stanza: A Python natural language processing toolkit for many human languages](https://md.research.elamur.ai/free/en/papers/qi-2020-stanza-python-natural-language-processing/). In Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics: System Demonstrations, pages 101–108, Online.…
14. Sunayana Sitaram, Khyathi Raghavi Chandu, Sai Krishna Rallabandi, and Alan W Black. 2020. A survey of code-switched speech and language processing. Preprint, arXiv:1904.00784. — [arXiv:1904.00784](https://arxiv.org/abs/1904.00784) [find](https://scholar.google.com/scholar?q=Sunayana+Sitaram%2C+Khyathi+Raghavi+Chandu%2C+Sai+Krishna+Rallabandi%2C+and+Alan+W+Black%2E+2020%2E+A+survey+of+code%2Dswitched+speech+and+language+processing%2E+Preprint%2C+arXiv%3A1904%2E00784%2E+%E2%80%94+%5BarXiv%3A1904%2E00784%5D%28htt)
15. Saba Sturua, Isabelle Mohr, Mohammad Kalim Akram, Michael Günther, Bo Wang, Markus Krimmel, Feng Wang, Georgios Mastrapas, Andreas Koukounas, Nan Wang, and Han Xiao. 2024. [jina-embeddings-v3: Multilingual embeddings with task LoRA](https://md.research.elamur.ai/free/en/papers/sturua-2024-jina-embeddings-v3-multilingual-embeddings/). Preprint, arXiv:2409.10173.
16. Liang Wang, Nan Yang, Xiaolong Huang, Linjun Yang, Rangan Majumder, and Furu Wei. 2024. [Multilingual E5 text embeddings: A technical report](https://md.research.elamur.ai/free/en/papers/wang-2024-multilingual-e5-text-embeddings-technical/). arXiv preprint arXiv:2402.05672.
17. Ruochen Zhang, Samuel Cahyawijaya, Jan Christian Blaise Cruz, Genta Winata, and Alham Fikri Aji. 2023a. [Multilingual large language models are not (yet) code-switchers](https://md.research.elamur.ai/free/en/papers/zhang-2023-multilingual-large-language-models-are/). In Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing, pages 12567–12582, Singapore. Associatio…
18. Xin Zhang, Yanzhao Zhang, Dingkun Long, Wen Xie, Ziqi Dai, Jialong Tang, Huan Lin, Baosong Yang, Pengjun Xie, Fei Huang, Meishan Zhang, Wenjie Li, and Min Zhang. 2024. [mGTE: Generalized long-context text representation and reranking models for multilingual text retrieval](https://md.research.elamur.ai/free/en/papers/zhang-2024-mgte-generalized-long-context-text/). In Proceedings of the 2024…
19. Xinyu Zhang, Xueguang Ma, Peng Shi, and Jimmy Lin. 2021. [Mr. TyDi: A multi-lingual benchmark for dense retrieval](https://md.research.elamur.ai/free/en/papers/zhang-2021-mr-tydi-multi-lingual-benchmark/). In Proceedings of the 1st Workshop on Multilingual Representation Learning, pages 127– 137, Punta Cana, Dominican Republic. Association for Computational Linguistics.
20. Xinyu Zhang, Nandan Thakur, Odunayo Ogundepo, Ehsan Kamalloo, David Alfonso-Hermelo, Xiaoguang Li, Qun Liu, Mehdi Rezagholizadeh, and Jimmy Lin. 2023b. [MIRACL: A multilingual retrieval dataset covering 18 diverse languages](https://md.research.elamur.ai/free/en/papers/zhang-2023-miracl-multilingual-retrieval-dataset-covering/). Transactions of the Association for Computational Linguistics, 11:1114–1131.
21. Yanzhao Zhang, Mingxin Li, Dingkun Long, Xin Zhang, Huan Lin, Baosong Yang, Pengjun Xie, An Yang, Dayiheng Liu, Junyang Lin, Fei Huang, and Jingren Zhou. 2025. [Qwen3 embedding: Advancing text embedding and reranking through foundation models](https://md.research.elamur.ai/free/en/papers/zhang-2025-qwen3-embedding-advancing-text/). arXiv preprint arXiv:2506.05176. do not override sampling…
22. A.4 *Figure 11: Illustration of Word-level query mixing via LLM, with examples.* [find](https://scholar.google.com/scholar?q=A%2E4+%2AFigure+11%3A+Illustration+of+Word%2Dlevel+query+mixing+via+LLM%2C+with+examples%2E%2A)

*Updated 2 November 2026*