Empowering low-resource languages: TraSe architecture for enhanced retrieval-augmented generation in Bangla
Open accessAbstract
Research on Retrieval-Augmented Generation for low-resource languages has been sparse because of limited resources.To address this, we focus on Bangla, a low-resource language, and have created a dataset of 200 question-answer pairs as a basis for our study from Bangla Wikipedia dump data.This paper introduces the TraSe architecture, which enhances RAG for Bangla using Translative prompting.Our experiments demonstrate that TraSe improves answer selection accuracy, achieving 34% with automatic retrieval and 63% with Human-in-the-Loop retrieval, outperforming baseline methods.The TraSe architecture marks a significant advancement in RAG for low-resource languages and has the potential to enhance question-answering systems for Bangla and similar languages.Future research could explore additional low-resource languages.
The claims and statistics of this paper haven’t been checked yet.
Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.
Integrity signals
Registry
Cites
Statistics
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
The TraSe architecture leverages cross-lingual translation to improve the generation of text-based outputs with improved synthesis.
“The TraSe architecture leverages cross-lingual translation to improve the generation of text-based outputs with improved synthesis (applied at the generation step).”
From Improving Amharic Information Retrieval with Translative and Multi-Agent Debate Retrieval Augmented Generation (2026) · cites this paper as Ipa et al., 2025
Section 3.4 says Translative prompting translates the Bangla query and context into English, finds an answer, and translates it back; the Results report better accuracy for text-based answers. But Sections 3.5 and 5 describe TraSe as selecting among already generated answers, not improving their synthesis at the generation step. (LLM, conf=0.85)
Experimental — accuracy not yet measuredTraSe circumvents native language deficits via translative prompting: translating query and retrieved context into English for generative processing, and translating final answer back to Bangla.
“TraSe circumvents native language deficits via”
From Improving Amharic Information Retrieval with Translative and Multi-Agent Debate Retrieval Augmented Generation (2026) · cites this paper as Ipa et al., 2025
TraSe uses a multi-prompt generation strategy coupled with an LLM-based selector to identify the most accurate candidate response.
“Furthermore, TraSe uses a multi-prompt generation strategy coupled with an LLM-based selector to identify the most accurate candidate response.”
From Improving Amharic Information Retrieval with Translative and Multi-Agent Debate Retrieval Augmented Generation (2026) · cites this paper as Ipa et al., 2025
Ipa et al. (2025) investigated RAG performance constraints in Bangla and introduced the TraSe architecture.
“A notable contribution addressing this gap is the work by (Ipa et al., 2025), who investigated RAG performance constraints in Bangla.Because standard models like Llama-2 struggle with native reasoning in underrepresented languages, they introduced the TraSe architecture.”
From Improving Amharic Information Retrieval with Translative and Multi-Agent Debate Retrieval Augmented Generation (2026) · cites this paper as Ipa et al., 2025
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
Cites
The publisher hasn’t deposited this paper’s reference list.
Research works printed but not linked: 3
- Abhik Bhattacharjee, Tahmid Hasan, Wasi Ah- mad, Kazi Samin Mubasshir, Md Saiful Is- lam, Anindya Iqbal, M. Sohel Rahman, and Rifat Shahriyar. 2022b. BanglaBERT: Lan- guage model pretraining and benchmarks for low-resource language understanding evaluation in Bangla. In Findings of the Association for Computational Linguistics: NAACL 2022, pages 1318-1327, Seattle, United States. Asso- ciation for Computational Linguistics. · DOI
- ES Shahul, Jithin James, Luis Espinosa Anke, and Steven Schockaert. 2023. Ragas: Automated evaluation of retrieval augmented generation. In Conference of the European Chapter of the As- sociation for Computational Linguistics.
- Fuda Ye, Shuangyin Li, Yongqi Zhang, and Lei Chen. 2024. R2ag: Incorporating retrieval in- formation into retrieval augmented generation. In Conference on Empirical Methods in Natural Language Processing.
Cited by
Sorted by citations
Jotie, A. A., & Mitra, P. (2026) · Proceedings of the Language Resources and Evaluation Conference · doi:10.63317/3hbm9z4okcrx
NextRAG: A Bilingual Retrieval-Augmented Generation Framework for Financial Question-Answering
Opin, S. A. I., Thakur, R. A., Raiyan, A. K., Kabir, M. R., & Rahman, S. M. A. (2025) · International Conference on Computer and Information Technology (ICCIT) · doi:10.1109/iccit68739.2025.11491522
Raiyan, M. S., & Ferdous, N. (2025) · International Conference on Computer and Information Technology (ICCIT) · doi:10.1109/iccit68739.2025.11491594
SEAR: Structured Reasoning Enhancement for Agricultural Rumor DetectionPreprint
Xiong, S., Si, Y., Hou, Y., Li, Y., Wang, Z., & Si, H. (2026) · SSRN Electronic Journal · doi:10.2139/ssrn.6881782
- Authors
- Ipa, A. S., Rony, M. A. T., & Islam, M. S.
- Affiliations
- Khulna University of Engineering and Technology (BD)
- Venue
- Proceedings of the 1st Workshop on Language Models for Underserved Communities (LM4UC 2025)
- Published
- 2025
- Pages
- 8-15
- OpenAlex
- W4411119997
- Cited by
- 3
- References
- 0
- Citations by year
- 2026: 1 · 2025: 2
- FWCI
- 2.84
- Type
- conference paper
- Language
- en
- Open access
- gold
- doi.org
- Licence
- cc-by
- Topics
- Physical Sciences › Computer Science › Artificial Intelligence › Topic Modeling
- Physical Sciences › Computer Science › Information Systems › Information Retrieval and Search Behavior
- Physical Sciences › Computer Science › Artificial Intelligence › Natural Language Processing Techniques
- Keywords
- retrieval-augmented generation, low-resource languages, Bengali language, question answering, answer selection, automatic retrieval, question answering dataset
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
