tirana-multilingual-bert
Query: multilingual BERT low resource language fine tuning
Results: 50
Date: 2026-07-08T07:34:56.939Z
1. Multilingual and Multimodal LLMs in the Wild: Building for Low-Resource Languages
Authors: Firoj Alam, Shammur Absar Chowdhury, Enamul Hoque Prince
Categories: cs.CL
Published: 2026-05-16
arXiv: 2605.17152v1
Abstract:
Multimodal LLMs are evolving from vision-language to tri-modality that see, hear, and read, yet pipelines and benchmarks remain English-centric and compute-heavy. The tutorial offers an overview of this emerging research area for multilingual multimodality across text, speech, and vision under limited data/compute budgets, synthesizing foundations, recent multilingual models (PALO, Maya), speech-text LLMs. We cover low-cost data creation/curation; adapter stacks for tri-modal alignment; culture-aware evaluation beyond English and hands on resources for fine-tuning a compact multilingual VLM and wiring a speech->text->LLM pipeline. The content will be delivered as an interactive half-day tutorial, designed for researchers and practitioners working on multilingual, multimodal AI in low-resource language settings.
2. Specializing Multilingual Language Models: An Empirical Study
Authors: Ethan C. Chau, Noah A. Smith
Categories: cs.CL
Published: 2021-06-16
arXiv: 2106.09063v4
Abstract:
Pretrained multilingual language models have become a common tool in transferring NLP capabilities to low-resource languages, often with adaptations. In this work, we study the performance, extensibility, and interaction of two such adaptations: vocabulary augmentation and script transliteration. Our evaluations on part-of-speech tagging, universal dependency parsing, and named entity recognition in nine diverse low-resource languages uphold the viability of these approaches while raising new questions around how to optimally adapt multilingual models to low-resource settings.
3. Multilingual Neural Machine Translation for Zero-Resource Languages
Authors: Surafel M. Lakew, Marcello Federico, Matteo Negri, Marco Turchi
Categories: cs.CL
Published: 2019-09-16
arXiv: 1909.07342v1
Abstract:
In recent years, Neural Machine Translation (NMT) has been shown to be more effective than phrase-based statistical methods, thus quickly becoming the state of the art in machine translation (MT). However, NMT systems are limited in translating low-resourced languages, due to the significant amount of parallel data that is required to learn useful mappings between languages. In this work, we show how the so-called multilingual NMT can help to tackle the challenges associated with low-resourced language translation. The underlying principle of multilingual NMT is to force the creation of hidden representations of words in a shared semantic space across multiple languages, thus enabling a positive parameter transfer across languages. Along this direction, we present multilingual translation experiments with three languages (English, Italian, Romanian) covering six translation directions, utilizing both recurrent neural networks and transformer (or self-attentive) neural networks. We then focus on the zero-shot translation problem, that is how to leverage multi-lingual data in order to learn translation directions that are not covered by the available training material. To this aim, we introduce our recently proposed iterative self-training method, which incrementally improves a multilingual NMT on a zero-shot direction by just relying on monolingual data. Our results on TED talks data show that multilingual NMT outperforms conventional bilingual NMT, that the transformer NMT outperforms recurrent NMT, and that zero-shot NMT outperforms conventional pivoting methods and even matches the performance of a fully-trained bilingual system.
4. Improving Multilingual Neural Machine Translation For Low-Resource Languages: French,English - Vietnamese
Authors: Thi-Vinh Ngo, Phuong-Thai Nguyen, Thanh-Le Ha, Khac-Quy Dinh, Le-Minh Nguyen
Categories: cs.CL, cs.LG
Published: 2020-12-16
arXiv: 2012.08743v2
Abstract:
Prior works have demonstrated that a low-resource language pair can benefit from multilingual machine translation (MT) systems, which rely on many language pairs’ joint training. This paper proposes two simple strategies to address the rare word issue in multilingual MT systems for two low-resource language pairs: French-Vietnamese and English-Vietnamese. The first strategy is about dynamical learning word similarity of tokens in the shared space among source languages while another one attempts to augment the translation ability of rare words through updating their embeddings during the training. Besides, we leverage monolingual data for multilingual MT systems to increase the amount of synthetic parallel corpora while dealing with the data sparsity problem. We have shown significant improvements of up to +1.62 and +2.54 BLEU points over the bilingual baseline systems for both language pairs and released our datasets for the research community.
5. Enhancing Plagiarism Detection in Marathi with a Weighted Ensemble of TF-IDF and BERT Embeddings for Low-Resource Language Processing
Authors: Atharva Mutsaddi, Aditya Choudhary
Categories: cs.CL, cs.AI, cs.LG
Published: 2025-01-09
arXiv: 2501.05260v1
Abstract:
Plagiarism involves using another person’s work or concepts without proper attribution, presenting them as original creations. With the growing amount of data communicated in regional languages such as Marathi – one of India’s regional languages – it is crucial to design robust plagiarism detection systems tailored for low-resource languages. Language models like Bidirectional Encoder Representations from Transformers (BERT) have demonstrated exceptional capability in text representation and feature extraction, making them essential tools for semantic analysis and plagiarism detection. However, the application of BERT for low-resource languages remains under-explored, particularly in the context of plagiarism detection. This paper presents a method to enhance the accuracy of plagiarism detection for Marathi texts using BERT sentence embeddings in conjunction with Term Frequency-Inverse Document Frequency (TF-IDF) feature representation. This approach effectively captures statistical, semantic, and syntactic aspects of text features through a weighted voting ensemble of machine learning models.
6. Overview of the First Workshop on Language Models for Low-Resource Languages (LoResLM 2025)
Authors: Hansi Hettiarachchi, Tharindu Ranasinghe, Paul Rayson, Ruslan Mitkov, Mohamed Gaber, Damith Premasiri, Fiona Anting Tan, Lasitha Uyangodage
Categories: cs.CL, cs.AI
Published: 2024-12-20
arXiv: 2412.16365v1
Abstract:
The first Workshop on Language Models for Low-Resource Languages (LoResLM 2025) was held in conjunction with the 31st International Conference on Computational Linguistics (COLING 2025) in Abu Dhabi, United Arab Emirates. This workshop mainly aimed to provide a forum for researchers to share and discuss their ongoing work on language models (LMs) focusing on low-resource languages, following the recent advancements in neural language models and their linguistic biases towards high-resource languages. LoResLM 2025 attracted notable interest from the natural language processing (NLP) community, resulting in 35 accepted papers from 52 submissions. These contributions cover a broad range of low-resource languages from eight language families and 13 diverse research areas, paving the way for future possibilities and promoting linguistic inclusivity in NLP.
7. Differentially Private Fine-tuning of Language Models
Authors: Da Yu, Saurabh Naik, Arturs Backurs, Sivakanth Gopi, Huseyin A. Inan, Gautam Kamath, Janardhan Kulkarni, Yin Tat Lee, Andre Manoel, Lukas Wutschitz, Sergey Yekhanin, Huishuai Zhang
Categories: cs.LG, cs.CL, cs.CR, stat.ML
Published: 2021-10-13
arXiv: 2110.06500v2
Abstract:
We give simpler, sparser, and faster algorithms for differentially private fine-tuning of large-scale pre-trained language models, which achieve the state-of-the-art privacy versus utility tradeoffs on many standard NLP tasks. We propose a meta-framework for this problem, inspired by the recent success of highly parameter-efficient methods for fine-tuning. Our experiments show that differentially private adaptations of these approaches outperform previous private algorithms in three important dimensions: utility, privacy, and the computational and memory cost of private training. On many commonly studied datasets, the utility of private models approaches that of non-private models. For example, on the MNLI dataset we achieve an accuracy of $87.8%$ using RoBERTa-Large and $83.5%$ using RoBERTa-Base with a privacy budget of $ε= 6.7$. In comparison, absent privacy constraints, RoBERTa-Large achieves an accuracy of $90.2%$. Our findings are similar for natural language generation tasks. Privately fine-tuning with DART, GPT-2-Small, GPT-2-Medium, GPT-2-Large, and GPT-2-XL achieve BLEU scores of 38.5, 42.0, 43.1, and 43.8 respectively (privacy budget of $ε= 6.8,δ=$ 1e-5) whereas the non-private baseline is $48.1$. All our experiments suggest that larger models are better suited for private fine-tuning: while they are well known to achieve superior accuracy non-privately, we find that they also better maintain their accuracy when privacy is introduced.
8. An Overview of Indian Spoken Language Recognition from Machine Learning Perspective
Authors: Spandan Dey, Md Sahidullah, Goutam Saha
Categories: cs.CL, cs.SD, eess.AS
Published: 2022-11-30
arXiv: 2212.03812v1
Abstract:
Automatic spoken language identification (LID) is a very important research field in the era of multilingual voice-command-based human-computer interaction (HCI). A front-end LID module helps to improve the performance of many speech-based applications in the multilingual scenario. India is a populous country with diverse cultures and languages. The majority of the Indian population needs to use their respective native languages for verbal interaction with machines. Therefore, the development of efficient Indian spoken language recognition systems is useful for adapting smart technologies in every section of Indian society. The field of Indian LID has started gaining momentum in the last two decades, mainly due to the development of several standard multilingual speech corpora for the Indian languages. Even though significant research progress has already been made in this field, to the best of our knowledge, there are not many attempts to analytically review them collectively. In this work, we have conducted one of the very first attempts to present a comprehensive review of the Indian spoken language recognition research field. In-depth analysis has been presented to emphasize the unique challenges of low-resource and mutual influences for developing LID systems in the Indian contexts. Several essential aspects of the Indian LID research, such as the detailed description of the available speech corpora, the major research contributions, including the earlier attempts based on statistical modeling to the recent approaches based on different neural network architectures, and the future research trends are discussed. This review work will help assess the state of the present Indian LID research by any active researcher or any research enthusiasts from related fields.
9. A Multilingual Sentiment Lexicon for Low-Resource Language Translation using Large Languages Models and Explainable AI
Authors: Melusi Malinga, Isaac Lupanda, Mike Wa Nkongolo, Phil van Deventer
Categories: cs.CL, cs.AI, cs.LG
Published: 2024-11-06
arXiv: 2411.04316v1
Abstract:
South Africa and the Democratic Republic of Congo (DRC) present a complex linguistic landscape with languages such as Zulu, Sepedi, Afrikaans, French, English, and Tshiluba (Ciluba), which creates unique challenges for AI-driven translation and sentiment analysis systems due to a lack of accurately labeled data. This study seeks to address these challenges by developing a multilingual lexicon designed for French and Tshiluba, now expanded to include translations in English, Afrikaans, Sepedi, and Zulu. The lexicon enhances cultural relevance in sentiment classification by integrating language-specific sentiment scores. A comprehensive testing corpus is created to support translation and sentiment analysis tasks, with machine learning models such as Random Forest, Support Vector Machine (SVM), Decision Trees, and Gaussian Naive Bayes (GNB) trained to predict sentiment across low resource languages (LRLs). Among them, the Random Forest model performed particularly well, capturing sentiment polarity and handling language-specific nuances effectively. Furthermore, Bidirectional Encoder Representations from Transformers (BERT), a Large Language Model (LLM), is applied to predict context-based sentiment with high accuracy, achieving 99% accuracy and 98% precision, outperforming other models. The BERT predictions were clarified using Explainable AI (XAI), improving transparency and fostering confidence in sentiment classification. Overall, findings demonstrate that the proposed lexicon and machine learning models significantly enhance translation and sentiment analysis for LRLs in South Africa and the DRC, laying a foundation for future AI models that support underrepresented languages, with applications across education, governance, and business in multilingual contexts.
10. On the Language-specificity of Multilingual BERT and the Impact of Fine-tuning
Authors: Marc Tanti, Lonneke van der Plas, Claudia Borg, Albert Gatt
Categories: cs.CL, cs.NE
Published: 2021-09-14
arXiv: 2109.06935v2
Abstract:
Recent work has shown evidence that the knowledge acquired by multilingual BERT (mBERT) has two components: a language-specific and a language-neutral one. This paper analyses the relationship between them, in the context of fine-tuning on two tasks – POS tagging and natural language inference – which require the model to bring to bear different degrees of language-specific knowledge. Visualisations reveal that mBERT loses the ability to cluster representations by language after fine-tuning, a result that is supported by evidence from language identification experiments. However, further experiments on ‘unlearning’ language-specific representations using gradient reversal and iterative adversarial learning are shown not to add further improvement to the language-independent component over and above the effect of fine-tuning. The results presented here suggest that the process of fine-tuning causes a reorganisation of the model’s limited representational capacity, enhancing language-independent representations at the expense of language-specific ones.
11. SPRING Lab IITM’s submission to Low Resource Indic Language Translation Shared Task
Authors: Hamees Sayed, Advait Joglekar, Srinivasan Umesh
Categories: cs.CL, cs.AI
Published: 2024-11-01
arXiv: 2411.00727v2
Abstract:
We develop a robust translation model for four low-resource Indic languages: Khasi, Mizo, Manipuri, and Assamese. Our approach includes a comprehensive pipeline from data collection and preprocessing to training and evaluation, leveraging data from WMT task datasets, BPCC, PMIndia, and OpenLanguageData. To address the scarcity of bilingual data, we use back-translation techniques on monolingual datasets for Mizo and Khasi, significantly expanding our training corpus. We fine-tune the pre-trained NLLB 3.3B model for Assamese, Mizo, and Manipuri, achieving improved performance over the baseline. For Khasi, which is not supported by the NLLB model, we introduce special tokens and train the model on our Khasi corpus. Our training involves masked language modelling, followed by fine-tuning for English-to-Indic and Indic-to-English translations.
12. Is Multilingual BERT Fluent in Language Generation?
Authors: Samuel Rönnqvist, Jenna Kanerva, Tapio Salakoski, Filip Ginter
Categories: cs.CL, cs.LG
Published: 2019-10-09
arXiv: 1910.03806v1
Abstract:
The multilingual BERT model is trained on 104 languages and meant to serve as a universal language model and tool for encoding sentences. We explore how well the model performs on several languages across several tasks: a diagnostic classification probing the embeddings for a particular syntactic property, a cloze task testing the language modelling ability to fill in gaps in a sentence, and a natural language generation task testing for the ability to produce coherent text fitting a given context. We find that the currently available multilingual BERT model is clearly inferior to the monolingual counterparts, and cannot in many cases serve as a substitute for a well-trained monolingual model. We find that the English and German models perform well at generation, whereas the multilingual model is lacking, in particular, for Nordic languages.
13. Demystifying Instruction Mixing for Fine-tuning Large Language Models
Authors: Renxi Wang, Haonan Li, Minghao Wu, Yuxia Wang, Xudong Han, Chiyu Zhang, Timothy Baldwin
Categories: cs.CL, cs.AI
Published: 2023-12-17
arXiv: 2312.10793v3
Abstract:
Instruction tuning significantly enhances the performance of large language models (LLMs) across various tasks. However, the procedure to optimizing the mixing of instruction datasets for LLM fine-tuning is still poorly understood. This study categorizes instructions into three primary types: NLP downstream tasks, coding, and general chat. We explore the effects of instruction tuning on different combinations of datasets on LLM performance, and find that certain instruction types are more advantageous for specific applications but can negatively impact other areas. This work provides insights into instruction mixtures, laying the foundations for future research.
14. Fine-tuning BERT for Low-Resource Natural Language Understanding via Active Learning
Authors: Daniel Grießhaber, Johannes Maucher, Ngoc Thang Vu
Categories: cs.CL
Published: 2020-12-04
arXiv: 2012.02462v1
Abstract:
Recently, leveraging pre-trained Transformer based language models in down stream, task specific models has advanced state of the art results in natural language understanding tasks. However, only a little research has explored the suitability of this approach in low resource settings with less than 1,000 training data points. In this work, we explore fine-tuning methods of BERT – a pre-trained Transformer based language model – by utilizing pool-based active learning to speed up training while keeping the cost of labeling new data constant. Our experimental results on the GLUE data set show an advantage in model performance by maximizing the approximate knowledge gain of the model when querying from the pool of unlabeled data. Finally, we demonstrate and analyze the benefits of freezing layers of the language model during fine-tuning to reduce the number of trainable parameters, making it more suitable for low-resource settings.
15. MMCR4NLP: Multilingual Multiway Corpora Repository for Natural Language Processing
Authors: Raj Dabre, Sadao Kurohashi
Categories: cs.CL
Published: 2017-10-03
arXiv: 1710.01025v3
Abstract:
Multilinguality is gradually becoming ubiquitous in the sense that more and more researchers have successfully shown that using additional languages help improve the results in many Natural Language Processing tasks. Multilingual Multiway Corpora (MMC) contain the same sentence in multiple languages. Such corpora have been primarily used for Multi-Source and Pivot Language Machine Translation but are also useful for developing multilingual sequence taggers by transfer learning. While these corpora are available, they are not organized for multilingual experiments and researchers need to write boilerplate code every time they want to use said corpora. Moreover, because there is no official MMC collection it becomes difficult to compare against existing approaches. As such we present our work on creating a unified and systematically organized repository of MMC spanning a large number of languages. We also provide training, development and test splits for corpora where official splits are unavailable. We hope that this will help speed up the pace of multilingual NLP research and ensure that NLP researchers obtain results that are more trustable since they can be compared easily. We indicate corpora sources, extraction procedures if any and relevant statistics. We also make our collection public for research purposes.
16. All Languages Matter: On the Multilingual Safety of Large Language Models
Authors: Wenxuan Wang, Zhaopeng Tu, Chang Chen, Youliang Yuan, Jen-tse Huang, Wenxiang Jiao, Michael R. Lyu
Categories: cs.CL, cs.AI
Published: 2023-10-02
arXiv: 2310.00905v2
Abstract:
Safety lies at the core of developing and deploying large language models (LLMs). However, previous safety benchmarks only concern the safety in one language, e.g. the majority language in the pretraining data such as English. In this work, we build the first multilingual safety benchmark for LLMs, XSafety, in response to the global deployment of LLMs in practice. XSafety covers 14 kinds of commonly used safety issues across 10 languages that span several language families. We utilize XSafety to empirically study the multilingual safety for 4 widely-used LLMs, including both close-API and open-source models. Experimental results show that all LLMs produce significantly more unsafe responses for non-English queries than English ones, indicating the necessity of developing safety alignment for non-English languages. In addition, we propose several simple and effective prompting methods to improve the multilingual safety of ChatGPT by evoking safety knowledge and improving cross-lingual generalization of safety alignment. Our prompting method can significantly reduce the ratio of unsafe responses from 19.1% to 9.7% for non-English queries. We release our data at https://github.com/Jarviswang94/Multilingual_safety_benchmark.
17. Automated Collection of Evaluation Dataset for Semantic Search in Low-Resource Domain Language
Authors: Anastasia Zhukova, Christian E. Matt, Bela Gipp
Categories: cs.CL
Published: 2024-12-13
arXiv: 2412.10008v1
Abstract:
Domain-specific languages that use a lot of specific terminology often fall into the category of low-resource languages. Collecting test datasets in a narrow domain is time-consuming and requires skilled human resources with domain knowledge and training for the annotation task. This study addresses the challenge of automated collecting test datasets to evaluate semantic search in low-resource domain-specific German language of the process industry. Our approach proposes an end-to-end annotation pipeline for automated query generation to the score reassessment of query-document pairs. To overcome the lack of text encoders trained in the German chemistry domain, we explore a principle of an ensemble of “weak” text encoders trained on common knowledge datasets. We combine individual relevance scores from diverse models to retrieve document candidates and relevance scores generated by an LLM, aiming to achieve consensus on query-document alignment. Evaluation results demonstrate that the ensemble method significantly improves alignment with human-assigned relevance scores, outperforming individual models in both inter-coder agreement and accuracy metrics. These findings suggest that ensemble learning can effectively adapt semantic search systems for specialized, low-resource languages, offering a practical solution to resource limitations in domain-specific contexts.
18. KenSwQuAD – A Question Answering Dataset for Swahili Low Resource Language
Authors: Barack W. Wanjawa, Lilian D. A. Wanzare, Florence Indede, Owen McOnyango, Lawrence Muchemi, Edward Ombui
Categories: cs.CL, cs.LG
Published: 2022-05-04
arXiv: 2205.02364v3
Abstract:
The need for Question Answering datasets in low resource languages is the motivation of this research, leading to the development of Kencorpus Swahili Question Answering Dataset, KenSwQuAD. This dataset is annotated from raw story texts of Swahili low resource language, which is a predominantly spoken in Eastern African and in other parts of the world. Question Answering (QA) datasets are important for machine comprehension of natural language for tasks such as internet search and dialog systems. Machine learning systems need training data such as the gold standard Question Answering set developed in this research. The research engaged annotators to formulate QA pairs from Swahili texts collected by the Kencorpus project, a Kenyan languages corpus. The project annotated 1,445 texts from the total 2,585 texts with at least 5 QA pairs each, resulting into a final dataset of 7,526 QA pairs. A quality assurance set of 12.5% of the annotated texts confirmed that the QA pairs were all correctly annotated. A proof of concept on applying the set to the QA task confirmed that the dataset can be usable for such tasks. KenSwQuAD has also contributed to resourcing of the Swahili language.
19. Dependency Annotation of Ottoman Turkish with Multilingual BERT
Authors: Şaziye Betül Özateş, Tarık Emre Tıraş, Efe Eren Genç, Esma Fatıma Bilgin Taşdemir
Categories: cs.CL
Published: 2024-02-22
arXiv: 2402.14743v2
Abstract:
This study introduces a pretrained large language model-based annotation methodology for the first de dency treebank in Ottoman Turkish. Our experimental results show that, iteratively, i) pseudo-annotating data using a multilingual BERT-based parsing model, ii) manually correcting the pseudo-annotations, and iii) fine-tuning the parsing model with the corrected annotations, we speed up and simplify the challenging dependency annotation process. The resulting treebank, that will be a part of the Universal Dependencies (UD) project, will facilitate automated analysis of Ottoman Turkish documents, unlocking the linguistic richness embedded in this historical heritage.
20. Revisiting Few-sample BERT Fine-tuning
Authors: Tianyi Zhang, Felix Wu, Arzoo Katiyar, Kilian Q. Weinberger, Yoav Artzi
Categories: cs.CL, cs.LG
Published: 2020-06-10
arXiv: 2006.05987v3
Abstract:
This paper is a study of fine-tuning of BERT contextual representations, with focus on commonly observed instabilities in few-sample scenarios. We identify several factors that cause this instability: the common use of a non-standard optimization method with biased gradient estimation; the limited applicability of significant parts of the BERT network for down-stream tasks; and the prevalent practice of using a pre-determined, and small number of training iterations. We empirically test the impact of these factors, and identify alternative practices that resolve the commonly observed instability of the process. In light of these observations, we re-visit recently proposed methods to improve few-sample fine-tuning with BERT and re-evaluate their effectiveness. Generally, we observe the impact of these methods diminishes significantly with our modified process.
21. UIO at SemEval-2023 Task 12: Multilingual fine-tuning for sentiment classification in low-resource languages
Authors: Egil Rønningstad
Categories: cs.CL
Published: 2023-04-27
arXiv: 2304.14189v1
Abstract:
Our contribution to the 2023 AfriSenti-SemEval shared task 12: Sentiment Analysis for African Languages, provides insight into how a multilingual large language model can be a resource for sentiment analysis in languages not seen during pretraining. The shared task provides datasets of a variety of African languages from different language families. The languages are to various degrees related to languages used during pretraining, and the language data contain various degrees of code-switching. We experiment with both monolingual and multilingual datasets for the final fine-tuning, and find that with the provided datasets that contain samples in the thousands, monolingual fine-tuning yields the best results.
22. Toward More Meaningful Resources for Lower-resourced Languages
Authors: Constantine Lignos, Nolan Holley, Chester Palen-Michel, Jonne Sälevä
Categories: cs.CL, cs.AI
Published: 2022-02-24
arXiv: 2202.12288v1
Abstract:
In this position paper, we describe our perspective on how meaningful resources for lower-resourced languages should be developed in connection with the speakers of those languages. We first examine two massively multilingual resources in detail. We explore the contents of the names stored in Wikidata for a few lower-resourced languages and find that many of them are not in fact in the languages they claim to be and require non-trivial effort to correct. We discuss quality issues present in WikiAnn and evaluate whether it is a useful supplement to hand annotated data. We then discuss the importance of creating annotation for lower-resourced languages in a thoughtful and ethical way that includes the languages’ speakers as part of the development process. We conclude with recommended guidelines for resource development.
23. adaptMLLM: Fine-Tuning Multilingual Language Models on Low-Resource Languages with Integrated LLM Playgrounds
Authors: Séamus Lankford, Haithem Afli, Andy Way
Categories: cs.CL, cs.AI
Published: 2024-03-04
arXiv: 2403.02370v1
Abstract:
The advent of Multilingual Language Models (MLLMs) and Large Language Models has spawned innovation in many areas of natural language processing. Despite the exciting potential of this technology, its impact on developing high-quality Machine Translation (MT) outputs for low-resource languages remains relatively under-explored. Furthermore, an open-source application, dedicated to both fine-tuning MLLMs and managing the complete MT workflow for low-resources languages, remains unavailable. We aim to address these imbalances through the development of adaptMLLM, which streamlines all processes involved in the fine-tuning of MLLMs for MT. This open-source application is tailored for developers, translators, and users who are engaged in MT. An intuitive interface allows for easy customisation of hyperparameters, and the application offers a range of metrics for model evaluation and the capability to deploy models as a translation service directly within the application. As a multilingual tool, we used adaptMLLM to fine-tune models for two low-resource language pairs: English to Irish (EN$\leftrightarrow$GA) and English to Marathi (EN$\leftrightarrow$MR). Compared with baselines from the LoResMT2021 Shared Task, the adaptMLLM system demonstrated significant improvements. In the EN$\rightarrow$GA direction, an improvement of 5.2 BLEU points was observed and an increase of 40.5 BLEU points was recorded in the GA$\rightarrow$EN direction. Significant improvements in the translation performance of the EN$\leftrightarrow$MR pair were also observed notably in the MR$\rightarrow$EN direction with an increase of 21.3 BLEU points. Finally, a fine-grained human evaluation of the MLLM output on the EN$\rightarrow$GA pair was conducted using the Multidimensional Quality Metrics and Scalar Quality Metrics error taxonomies. The application and models are freely available.
24. BERT Based Multilingual Machine Comprehension in English and Hindi
Authors: Somil Gupta, Nilesh Khade
Categories: cs.CL
Published: 2020-06-02
arXiv: 2006.01432v1
Abstract:
Multilingual Machine Comprehension (MMC) is a Question-Answering (QA) sub-task that involves quoting the answer for a question from a given snippet, where the question and the snippet can be in different languages. Recently released multilingual variant of BERT (m-BERT), pre-trained with 104 languages, has performed well in both zero-shot and fine-tuned settings for multilingual tasks; however, it has not been used for English-Hindi MMC yet. We, therefore, present in this article, our experiments with m-BERT for MMC in zero-shot, mono-lingual (e.g. Hindi Question-Hindi Snippet) and cross-lingual (e.g. English QuestionHindi Snippet) fine-tune setups. These model variants are evaluated on all possible multilingual settings and results are compared against the current state-of-the-art sequential QA system for these languages. Experiments show that m-BERT, with fine-tuning, improves performance on all evaluation settings across both the datasets used by the prior model, therefore establishing m-BERT based MMC as the new state-of-the-art for English and Hindi. We also publish our results on an extended version of the recently released XQuAD dataset, which we propose to use as the evaluation benchmark for future research.
25. Cross-lingual transfer of multilingual models on low resource African Languages
Authors: Harish Thangaraj, Ananya Chenat, Jaskaran Singh Walia, Vukosi Marivate
Categories: cs.CL, cs.LG
Published: 2024-09-17
arXiv: 2409.10965v1
Abstract:
Large multilingual models have significantly advanced natural language processing (NLP) research. However, their high resource demands and potential biases from diverse data sources have raised concerns about their effectiveness across low-resource languages. In contrast, monolingual models, trained on a single language, may better capture the nuances of the target language, potentially providing more accurate results. This study benchmarks the cross-lingual transfer capabilities from a high-resource language to a low-resource language for both, monolingual and multilingual models, focusing on Kinyarwanda and Kirundi, two Bantu languages. We evaluate the performance of transformer based architectures like Multilingual BERT (mBERT), AfriBERT, and BantuBERTa against neural-based architectures such as BiGRU, CNN, and char-CNN. The models were trained on Kinyarwanda and tested on Kirundi, with fine-tuning applied to assess the extent of performance improvement and catastrophic forgetting. AfriBERT achieved the highest cross-lingual accuracy of 88.3% after fine-tuning, while BiGRU emerged as the best-performing neural model with 83.3% accuracy. We also analyze the degree of forgetting in the original language post-fine-tuning. While monolingual models remain competitive, this study highlights that multilingual models offer strong cross-lingual transfer capabilities in resource limited settings.
26. Speechless: Speech Instruction Training Without Speech for Low Resource Languages
Authors: Alan Dao, Dinh Bach Vu, Huy Hoang Ha, Tuan Le Duc Anh, Shreyas Gopal, Yue Heng Yeo, Warren Keng Hoong Low, Eng Siong Chng, Jia Qi Yip
Categories: eess.AS, cs.CL, cs.SD
Published: 2025-05-23
arXiv: 2505.17417v1
Abstract:
The rapid growth of voice assistants powered by large language models (LLM) has highlighted a need for speech instruction data to train these systems. Despite the abundance of speech recognition data, there is a notable scarcity of speech instruction data, which is essential for fine-tuning models to understand and execute spoken commands. Generating high-quality synthetic speech requires a good text-to-speech (TTS) model, which may not be available to low resource languages. Our novel approach addresses this challenge by halting synthesis at the semantic representation level, bypassing the need for TTS. We achieve this by aligning synthetic semantic representations with the pre-trained Whisper encoder, enabling an LLM to be fine-tuned on text instructions while maintaining the ability to understand spoken instructions during inference. This simplified training process is a promising approach to building voice assistant for low-resource languages.
27. Challenges in Adapting Multilingual LLMs to Low-Resource Languages using LoRA PEFT Tuning
Authors: Omkar Khade, Shruti Jagdale, Abhishek Phaltankar, Gauri Takalikar, Raviraj Joshi
Categories: cs.CL, cs.LG
Published: 2024-11-27
arXiv: 2411.18571v1
Abstract:
Large Language Models (LLMs) have demonstrated remarkable multilingual capabilities, yet challenges persist in adapting these models for low-resource languages. In this study, we investigate the effects of Low-Rank Adaptation (LoRA) Parameter-Efficient Fine-Tuning (PEFT) on multilingual Gemma models for Marathi, a language with limited resources. Using a translated Alpaca dataset with 52,000 instruction-response pairs, our findings reveal that while evaluation metrics often show a performance decline post-fine-tuning, manual assessments frequently suggest that the fine-tuned models outperform their original counterparts. The observations indicate improvements in target language generation capabilities but a reduction in reasoning abilities following language adaptation. These results underscore the need for improved evaluation methodologies and the creation of high-quality native datasets to accurately assess language-specific model performance in low-resource settings.
28. Multilingual Offensive Language Identification for Low-resource Languages
Authors: Tharindu Ranasinghe, Marcos Zampieri
Categories: cs.CL, cs.AI, cs.LG, cs.SI
Published: 2021-05-12
arXiv: 2105.05996v3
Abstract:
Offensive content is pervasive in social media and a reason for concern to companies and government organizations. Several studies have been recently published investigating methods to detect the various forms of such content (e.g. hate speech, cyberbullying, and cyberaggression). The clear majority of these studies deal with English partially because most annotated datasets available contain English data. In this paper, we take advantage of available English datasets by applying cross-lingual contextual word embeddings and transfer learning to make predictions in low-resource languages. We project predictions on comparable data in Arabic, Bengali, Danish, Greek, Hindi, Spanish, and Turkish. We report results of 0.8415 F1 macro for Bengali in TRAC-2 shared task, 0.8532 F1 macro for Danish and 0.8701 F1 macro for Greek in OffensEval 2020, 0.8568 F1 macro for Hindi in HASOC 2019 shared task and 0.7513 F1 macro for Spanish in in SemEval-2019 Task 5 (HatEval) showing that our approach compares favourably to the best systems submitted to recent shared tasks on these three languages. Additionally, we report competitive performance on Arabic, and Turkish using the training and development sets of OffensEval 2020 shared task. The results for all languages confirm the robustness of cross-lingual contextual embeddings and transfer learning for this task.
29. Extending Multilingual BERT to Low-Resource Languages
Authors: Zihan Wang, Karthikeyan K, Stephen Mayhew, Dan Roth
Categories: cs.CL
Published: 2020-04-28
arXiv: 2004.13640v1
Abstract:
Multilingual BERT (M-BERT) has been a huge success in both supervised and zero-shot cross-lingual transfer learning. However, this success has focused only on the top 104 languages in Wikipedia that it was trained on. In this paper, we propose a simple but effective approach to extend M-BERT (E-BERT) so that it can benefit any new language, and show that our approach benefits languages that are already in M-BERT as well. We perform an extensive set of experiments with Named Entity Recognition (NER) on 27 languages, only 16 of which are in M-BERT, and show an average increase of about 6% F1 on languages that are already in M-BERT and 23% F1 increase on new languages.
30. Role of Language Relatedness in Multilingual Fine-tuning of Language Models: A Case Study in Indo-Aryan Languages
Authors: Tejas Indulal Dhamecha, Rudra Murthy, Samarth Bharadwaj, Karthik Sankaranarayanan, Pushpak Bhattacharyya
Categories: cs.CL
Published: 2021-09-22
arXiv: 2109.10534v1
Abstract:
We explore the impact of leveraging the relatedness of languages that belong to the same family in NLP models using multilingual fine-tuning. We hypothesize and validate that multilingual fine-tuning of pre-trained language models can yield better performance on downstream NLP applications, compared to models fine-tuned on individual languages. A first of its kind detailed study is presented to track performance change as languages are added to a base language in a graded and greedy (in the sense of best boost of performance) manner; which reveals that careful selection of subset of related languages can significantly improve performance than utilizing all related languages. The Indo-Aryan (IA) language family is chosen for the study, the exact languages being Bengali, Gujarati, Hindi, Marathi, Oriya, Punjabi and Urdu. The script barrier is crossed by simple rule-based transliteration of the text of all languages to Devanagari. Experiments are performed on mBERT, IndicBERT, MuRIL and two RoBERTa-based LMs, the last two being pre-trained by us. Low resource languages, such as Oriya and Punjabi, are found to be the largest beneficiaries of multilingual fine-tuning. Textual Entailment, Entity Classification, Section Title Prediction, tasks of IndicGLUE and POS tagging form our test bed. Compared to monolingual fine tuning we get relative performance improvement of up to 150% in the downstream tasks. The surprise take-away is that for any language there is a particular combination of other languages which yields the best performance, and any additional language is in fact detrimental.
31. Breaking Physical and Linguistic Borders: Multilingual Federated Prompt Tuning for Low-Resource Languages
Authors: Wanru Zhao, Yihong Chen, Royson Lee, Xinchi Qiu, Yan Gao, Hongxiang Fan, Nicholas D. Lane
Categories: cs.CL
Published: 2025-07-02
arXiv: 2507.03003v1
Abstract:
Pre-trained large language models (LLMs) have become a cornerstone of modern natural language processing, with their capabilities extending across a wide range of applications and languages. However, the fine-tuning of multilingual LLMs, especially for low-resource languages, faces significant challenges arising from data-sharing restrictions (the physical border) and inherent linguistic differences (the linguistic border). These barriers hinder users of various languages, particularly those in low-resource regions, from fully benefiting from the advantages of LLMs. To address these challenges, we propose the Federated Prompt Tuning Paradigm for multilingual scenarios, which utilizes parameter-efficient fine-tuning while adhering to data sharing restrictions. We design a comprehensive set of experiments and analyze them using a novel notion of language distance to highlight the strengths of our paradigm: Even under computational constraints, our method not only improves data efficiency but also facilitates mutual enhancements across languages, particularly benefiting low-resource ones. Compared to traditional local cross-lingual transfer tuning methods, our approach achieves 6.9% higher accuracy with improved data efficiency, and demonstrates greater stability and generalization. These findings underscore the potential of our approach to promote social equality and champion linguistic diversity, ensuring that no language is left behind.
32. Massively Multilingual Text Translation For Low-Resource Languages
Authors: Zhong Zhou
Categories: cs.CL
Published: 2024-01-29
arXiv: 2401.16582v1
Abstract:
Translation into severely low-resource languages has both the cultural goal of saving and reviving those languages and the humanitarian goal of assisting the everyday needs of local communities that are accelerated by the recent COVID-19 pandemic. In many humanitarian efforts, translation into severely low-resource languages often does not require a universal translation engine, but a dedicated text-specific translation engine. For example, healthcare records, hygienic procedures, government communication, emergency procedures and religious texts are all limited texts. While generic translation engines for all languages do not exist, translation of multilingually known limited texts into new, low-resource languages may be possible and reduce human translation effort. We attempt to leverage translation resources from rich-resource languages to efficiently produce best possible translation quality for well known texts, which are available in multiple languages, in a new, low-resource language. To reach this goal, we argue that in translating a closed text into low-resource languages, generalization to out-of-domain texts is not necessary, but generalization to new languages is. Performance gain comes from massive source parallelism by careful choice of close-by language families, style-consistent corpus-level paraphrases within the same language and strategic adaptation of existing large pretrained multilingual models to the domain first and then to the language. Such performance gain makes it possible for machine translation systems to collaborate with human translators to expedite the translation process into new, low-resource languages.
33. Building a Strong Instruction Language Model for a Less-Resourced Language
Authors: Domen Vreš, Tjaša Arčon, Timotej Petrič, Dario Vajda, Marko Robnik-Šikonja, Iztok Lebar Bajec
Categories: cs.CL, cs.LG
Published: 2026-03-02
arXiv: 2603.01691v1
Abstract:
Large language models (LLMs) have become an essential tool for natural language processing and artificial intelligence in general. Current open-source models are primarily trained on English texts, resulting in poorer performance on less-resourced languages and cultures. We present a set of methodological approaches necessary for the successful adaptation of an LLM to a less-resourced language, and demonstrate them using the Slovene language. We present GaMS3-12B, a generative model for Slovene with 12 billion parameters, and demonstrate that it is the best-performing open-source model for Slovene within its parameter range. We adapted the model to the Slovene language using three-stage continual pre-training of the Gemma 3 model, followed by two-stage supervised fine-tuning (SFT). We trained the model on a combination of 140B Slovene, English, Bosnian, Serbian, and Croatian pretraining tokens, and over 200 thousand English and Slovene SFT examples. We evaluate GaMS3-12B on the Slovenian-LLM-Eval datasets, English-to-Slovene translation, and the Slovene LLM arena. We show that the described model outperforms 12B Gemma 3 across all three scenarios and performs comparably to much larger commercial GPT-4o in the Slovene LLM arena, achieving a win rate of over 60 %.
34. MyCulture: Exploring Malaysia’s Diverse Culture under Low-Resource Language Constraints
Authors: Zhong Ken Hew, Jia Xin Low, Sze Jue Yang, Chee Seng Chan
Categories: cs.CL, cs.AI
Published: 2025-08-07
arXiv: 2508.05429v2
Abstract:
Large Language Models (LLMs) often exhibit cultural biases due to training data dominated by high-resource languages like English and Chinese. This poses challenges for accurately representing and evaluating diverse cultural contexts, particularly in low-resource language settings. To address this, we introduce MyCulture, a benchmark designed to comprehensively evaluate LLMs on Malaysian culture across six pillars: arts, attire, customs, entertainment, food, and religion presented in Bahasa Melayu. Unlike conventional benchmarks, MyCulture employs a novel open-ended multiple-choice question format without predefined options, thereby reducing guessing and mitigating format bias. We provide a theoretical justification for the effectiveness of this open-ended structure in improving both fairness and discriminative power. Furthermore, we analyze structural bias by comparing model performance on structured versus free-form outputs, and assess language bias through multilingual prompt variations. Our evaluation across a range of regional and international LLMs reveals significant disparities in cultural comprehension, highlighting the urgent need for culturally grounded and linguistically inclusive benchmarks in the development and assessment of LLMs.
35. DN at SemEval-2023 Task 12: Low-Resource Language Text Classification via Multilingual Pretrained Language Model Fine-tuning
Authors: Daniil Homskiy, Narek Maloyan
Categories: cs.CL
Published: 2023-05-04
arXiv: 2305.02607v1
Abstract:
In recent years, sentiment analysis has gained significant importance in natural language processing. However, most existing models and datasets for sentiment analysis are developed for high-resource languages, such as English and Chinese, leaving low-resource languages, particularly African languages, largely unexplored. The AfriSenti-SemEval 2023 Shared Task 12 aims to fill this gap by evaluating sentiment analysis models on low-resource African languages. In this paper, we present our solution to the shared task, where we employed different multilingual XLM-R models with classification head trained on various data, including those retrained in African dialects and fine-tuned on target languages. Our team achieved the third-best results in Subtask B, Track 16: Multilingual, demonstrating the effectiveness of our approach. While our model showed relatively good results on multilingual data, it performed poorly in some languages. Our findings highlight the importance of developing more comprehensive datasets and models for low-resource African languages to advance sentiment analysis research. We also provided the solution on the github repository.
36. Is The Universal Matter - Antimatter Asymmetry Fine Tuned?
Authors: Gary Steigman, Robert J. Scherrer
Categories: astro-ph.CO, hep-ph
Published: 2018-01-30
arXiv: 1801.10059v1
Abstract:
The asymmetry between matter and antimatter is key to the existence and nature of our Universe. A measure of the matter - antimatter asymmetry of the Universe is provided by the present value of the universal ratio of baryons (baryons minus antibaryons) to photons (or the ratio of baryons to entropy). The baryon asymmetry parameter is an important physical and cosmological parameter. But how fine tuned is it? A “natural” value for this parameter is zero, corresponding to equal amounts of matter and antimatter. Another, also possibly natural, choice for this dimensionless parameter would be of order unity, corresponding to nearly equal amounts (by number) of matter (and essentially no antimatter) and photons in every comoving volume. However, observations suggest that in the Universe we inhabit the value of this parameter is nonzero, but smaller than this natural value by some nine to ten orders of magnitude. In this contribution we review the evidence that our Universe does not contain equal amounts of matter and antimatter. Any change in the magnitude of the baryon asymmetry parameter necessarily leads to a universe with physical characteristics different from those in our own. The degree of fine tuning in the baryon asymmetry parameter is determined by the width of the range over which it can be varied and still allow for the existence of life. Our results suggest that the baryon asymmetry parameter can be varied over a very wide range without impacting the prospects for life; this result is not suggestive of fine tuning. [abridged]
37. Beyond Data Quantity: Key Factors Driving Performance in Multilingual Language Models
Authors: Sina Bagheri Nezhad, Ameeta Agrawal, Rhitabrat Pokharel
Categories: cs.CL, cs.AI
Published: 2024-12-17
arXiv: 2412.12500v1
Abstract:
Multilingual language models (MLLMs) are crucial for handling text across various languages, yet they often show performance disparities due to differences in resource availability and linguistic characteristics. While the impact of pre-train data percentage and model size on performance is well-known, our study reveals additional critical factors that significantly influence MLLM effectiveness. Analyzing a wide range of features, including geographical, linguistic, and resource-related aspects, we focus on the SIB-200 dataset for classification and the Flores-200 dataset for machine translation, using regression models and SHAP values across 204 languages. Our findings identify token similarity and country similarity as pivotal factors, alongside pre-train data and model size, in enhancing model performance. Token similarity facilitates cross-lingual transfer, while country similarity highlights the importance of shared cultural and linguistic contexts. These insights offer valuable guidance for developing more equitable and effective multilingual language models, particularly for underrepresented languages.
38. The Road to Know-Where: An Object-and-Room Informed Sequential BERT for Indoor Vision-Language Navigation
Authors: Yuankai Qi, Zizheng Pan, Yicong Hong, Ming-Hsuan Yang, Anton van den Hengel, Qi Wu
Categories: cs.CL, cs.CV
Published: 2021-04-09
arXiv: 2104.04167v2
Abstract:
Vision-and-Language Navigation (VLN) requires an agent to find a path to a remote location on the basis of natural-language instructions and a set of photo-realistic panoramas. Most existing methods take the words in the instructions and the discrete views of each panorama as the minimal unit of encoding. However, this requires a model to match different nouns (e.g., TV, table) against the same input view feature. In this work, we propose an object-informed sequential BERT to encode visual perceptions and linguistic instructions at the same fine-grained level, namely objects and words. Our sequential BERT also enables the visual-textual clues to be interpreted in light of the temporal context, which is crucial to multi-round VLN tasks. Additionally, we enable the model to identify the relative direction (e.g., left/right/front/back) of each navigable location and the room type (e.g., bedroom, kitchen) of its current and final navigation goal, as such information is widely mentioned in instructions implying the desired next and final locations. We thus enable the model to know-where the objects lie in the images, and to know-where they stand in the scene. Extensive experiments demonstrate the effectiveness compared against several state-of-the-art methods on three indoor VLN tasks: REVERIE, NDH, and R2R. Project repository: https://github.com/YuankaiQi/ORIST
39. Are All Languages Created Equal in Multilingual BERT?
Authors: Shijie Wu, Mark Dredze
Categories: cs.CL
Published: 2020-05-18
arXiv: 2005.09093v2
Abstract:
Multilingual BERT (mBERT) trained on 104 languages has shown surprisingly good cross-lingual performance on several NLP tasks, even without explicit cross-lingual signals. However, these evaluations have focused on cross-lingual transfer with high-resource languages, covering only a third of the languages covered by mBERT. We explore how mBERT performs on a much wider set of languages, focusing on the quality of representation for low-resource languages, measured by within-language performance. We consider three tasks: Named Entity Recognition (99 languages), Part-of-speech Tagging, and Dependency Parsing (54 languages each). mBERT does better than or comparable to baselines on high resource languages but does much worse for low resource languages. Furthermore, monolingual BERT models for these languages do even worse. Paired with similar languages, the performance gap between monolingual BERT and mBERT can be narrowed. We find that better models for low resource languages require more efficient pretraining techniques or more data.
40. Filipino Benchmarks for Measuring Sexist and Homophobic Bias in Multilingual Language Models from Southeast Asia
Authors: Lance Calvin Lim Gamboa, Mark Lee
Categories: cs.CL
Published: 2024-12-10
arXiv: 2412.07303v2
Abstract:
Bias studies on multilingual models confirm the presence of gender-related stereotypes in masked models processing languages with high NLP resources. We expand on this line of research by introducing Filipino CrowS-Pairs and Filipino WinoQueer: benchmarks that assess both sexist and anti-queer biases in pretrained language models (PLMs) handling texts in Filipino, a low-resource language from the Philippines. The benchmarks consist of 7,074 new challenge pairs resulting from our cultural adaptation of English bias evaluation datasets, a process that we document in detail to guide similar forthcoming efforts. We apply the Filipino benchmarks on masked and causal multilingual models, including those pretrained on Southeast Asian data, and find that they contain considerable amounts of bias. We also find that for multilingual models, the extent of bias learned for a particular language is influenced by how much pretraining data in that language a model was exposed to. Our benchmarks and insights can serve as a foundation for future work analyzing and mitigating bias in multilingual models.
41. Weighted Cross-entropy for Low-Resource Languages in Multilingual Speech Recognition
Authors: Andrés Piñeiro-Martín, Carmen García-Mateo, Laura Docío-Fernández, María del Carmen López-Pérez, Georg Rehm
Categories: cs.CL, cs.SD, eess.AS
Published: 2024-09-25
arXiv: 2409.16954v1
Abstract:
This paper addresses the challenge of integrating low-resource languages into multilingual automatic speech recognition (ASR) systems. We introduce a novel application of weighted cross-entropy, typically used for unbalanced datasets, to facilitate the integration of low-resource languages into pre-trained multilingual ASR models within the context of continual multilingual learning. We fine-tune the Whisper multilingual ASR model on five high-resource languages and one low-resource language, employing language-weighted dynamic cross-entropy and data augmentation. The results show a remarkable 6.69% word error rate (WER) reduction for the low-resource language compared to the fine-tuned model without applying our approach, and a 48.86% WER reduction compared to the original Whisper model. In addition, our approach yields an average WER reduction of 3.29% across the six languages, showing no degradation for the high-resource languages.
42. Improved acoustic word embeddings for zero-resource languages using multilingual transfer
Authors: Herman Kamper, Yevgen Matusevych, Sharon Goldwater
Categories: cs.CL, cs.SD, eess.AS
Published: 2020-06-02
arXiv: 2006.02295v2
Abstract:
Acoustic word embeddings are fixed-dimensional representations of variable-length speech segments. Such embeddings can form the basis for speech search, indexing and discovery systems when conventional speech recognition is not possible. In zero-resource settings where unlabelled speech is the only available resource, we need a method that gives robust embeddings on an arbitrary language. Here we explore multilingual transfer: we train a single supervised embedding model on labelled data from multiple well-resourced languages and then apply it to unseen zero-resource languages. We consider three multilingual recurrent neural network (RNN) models: a classifier trained on the joint vocabularies of all training languages; a Siamese RNN trained to discriminate between same and different words from multiple languages; and a correspondence autoencoder (CAE) RNN trained to reconstruct word pairs. In a word discrimination task on six target languages, all of these models outperform state-of-the-art unsupervised models trained on the zero-resource languages themselves, giving relative improvements of more than 30% in average precision. When using only a few training languages, the multilingual CAE performs better, but with more training languages the other multilingual models perform similarly. Using more training languages is generally beneficial, but improvements are marginal on some languages. We present probing experiments which show that the CAE encodes more phonetic, word duration, language identity and speaker information than the other multilingual models.
43. Evaluating Standard and Dialectal Frisian ASR: Multilingual Fine-tuning and Language Identification for Improved Low-resource Performance
Authors: Reihaneh Amooie, Wietse de Vries, Yun Hao, Jelske Dijkstra, Matt Coler, Martijn Wieling
Categories: cs.CL, cs.LG, cs.SD, eess.AS
Published: 2025-02-07
arXiv: 2502.04883v1
Abstract:
Automatic Speech Recognition (ASR) performance for low-resource languages is still far behind that of higher-resource languages such as English, due to a lack of sufficient labeled data. State-of-the-art methods deploy self-supervised transfer learning where a model pre-trained on large amounts of data is fine-tuned using little labeled data in a target low-resource language. In this paper, we present and examine a method for fine-tuning an SSL-based model in order to improve the performance for Frisian and its regional dialects (Clay Frisian, Wood Frisian, and South Frisian). We show that Frisian ASR performance can be improved by using multilingual (Frisian, Dutch, English and German) fine-tuning data and an auxiliary language identification task. In addition, our findings show that performance on dialectal speech suffers substantially, and, importantly, that this effect is moderated by the elicitation approach used to collect the dialectal data. Our findings also particularly suggest that relying solely on standard language data for ASR evaluation may underestimate real-world performance, particularly in languages with substantial dialectal variation.
44. SpokenNativQA: Multilingual Everyday Spoken Queries for LLMs
Authors: Firoj Alam, Md Arid Hasan, Shammur Absar Chowdhury
Categories: cs.CL, cs.AI
Published: 2025-05-25
arXiv: 2505.19163v1
Abstract:
Large Language Models (LLMs) have demonstrated remarkable performance across various disciplines and tasks. However, benchmarking their capabilities with multilingual spoken queries remains largely unexplored. In this study, we introduce SpokenNativQA, the first multilingual and culturally aligned spoken question-answering (SQA) dataset designed to evaluate LLMs in real-world conversational settings. The dataset comprises approximately 33,000 naturally spoken questions and answers in multiple languages, including low-resource and dialect-rich languages, providing a robust benchmark for assessing LLM performance in speech-based interactions. SpokenNativQA addresses the limitations of text-based QA datasets by incorporating speech variability, accents, and linguistic diversity. We benchmark different ASR systems and LLMs for SQA and present our findings. We released the data at (https://huggingface.co/datasets/QCRI/SpokenNativQA) and the experimental scripts at (https://llmebench.qcri.org/) for the research community.
45. End-to-End Spoken Language Understanding: Performance analyses of a voice command task in a low resource setting
Authors: Thierry Desot, François Portet, Michel Vacher
Categories: cs.CL, cs.SD, eess.AS
Published: 2022-07-17
arXiv: 2207.08179v1
Abstract:
Spoken Language Understanding (SLU) is a core task in most human-machine interaction systems. With the emergence of smart homes, smart phones and smart speakers, SLU has become a key technology for the industry. In a classical SLU approach, an Automatic Speech Recognition (ASR) module transcribes the speech signal into a textual representation from which a Natural Language Understanding (NLU) module extracts semantic information. Recently End-to-End SLU (E2E SLU) based on Deep Neural Networks has gained momentum since it benefits from the joint optimization of the ASR and the NLU parts, hence limiting the cascade of error effect of the pipeline architecture. However, little is known about the actual linguistic properties used by E2E models to predict concepts and intents from speech input. In this paper, we present a study identifying the signal features and other linguistic properties used by an E2E model to perform the SLU task. The study is carried out in the application domain of a smart home that has to handle non-English (here French) voice commands. The results show that a good E2E SLU performance does not always require a perfect ASR capability. Furthermore, the results show the superior capabilities of the E2E model in handling background noise and syntactic variation compared to the pipeline model. Finally, a finer-grained analysis suggests that the E2E model uses the pitch information of the input signal to identify voice command concepts. The results and methodology outlined in this paper provide a springboard for further analyses of E2E models in speech processing.
46. UR-BERT: Scaling Text Encoders for Massively Multilingual TTS Through Universal Romanization and Speech Token Prediction
Authors: Sangmin Lee, Eekgyun Ahn, Woongjib Choi, Hong-Goo Kang
Categories: cs.CL, cs.SD
Published: 2026-06-10
arXiv: 2606.11681v2
Abstract:
We propose UR-BERT, a Romanized transcription-based text-to-speech (TTS) encoder for massively multilingual TTS systems. Conventional grapheme-to-phoneme (G2P)-based approaches are limited to around 100 languages due to the availability of reliable G2P resources. In contrast, UR-BERT scales to 495 languages by unifying diverse writing systems into a shared Romanization representation. To further enhance phonetic fidelity and text-speech alignment, we introduce a speech token prediction objective during training, which encourages the encoder to learn speech-aware phonetic representations in a data-efficient manner. Experiments show that TTS systems built on UR-BERT consistently outperform recent text encoder baselines across a wide range of languages and resource conditions, and demonstrate strong generalization to unseen languages.
47. Multilingual Transformer Language Model for Speech Recognition in Low-resource Languages
Authors: Li Miao, Jian Wu, Piyush Behre, Shuangyu Chang, Sarangarajan Parthasarathy
Categories: cs.CL
Published: 2022-09-08
arXiv: 2209.04041v1
Abstract:
It is challenging to train and deploy Transformer LMs for hybrid speech recognition 2nd pass re-ranking in low-resource languages due to (1) data scarcity in low-resource languages, (2) expensive computing costs for training and refreshing 100+ monolingual models, and (3) hosting inefficiency considering sparse traffic. In this study, we present a new way to group multiple low-resource locales together and optimize the performance of Multilingual Transformer LMs in ASR. Our Locale-group Multilingual Transformer LMs outperform traditional multilingual LMs along with reducing maintenance costs and operating expenses. Further, for low-resource but high-traffic locales where deploying monolingual models is feasible, we show that fine-tuning our locale-group multilingual LMs produces better monolingual LM candidates than baseline monolingual LMs.
48. Exploiting multilingual nomenclatures and language-independent text features as an interlingua for cross-lingual text analysis applications
Authors: Ralf Steinberger, Bruno Pouliquen, Camelia Ignat
Categories: cs.CL, cs.IR
Published: 2006-09-12
arXiv: cs/0609064v1
Abstract:
We are proposing a simple, but efficient basic approach for a number of multilingual and cross-lingual language technology applications that are not limited to the usual two or three languages, but that can be applied with relatively little effort to larger sets of languages. The approach consists of using existing multilingual linguistic resources such as thesauri, nomenclatures and gazetteers, as well as exploiting the existence of additional more or less language-independent text items such as dates, currency expressions, numbers, names and cognates. Mapping texts onto the multilingual resources and identifying word token links between texts in different languages are basic ingredients for applications such as cross-lingual document similarity calculation, multilingual clustering and categorisation, cross-lingual document retrieval, and tools to provide cross-lingual information access.
49. Adapting Pre-trained Language Models to African Languages via Multilingual Adaptive Fine-Tuning
Authors: Jesujoba O. Alabi, David Ifeoluwa Adelani, Marius Mosbach, Dietrich Klakow
Categories: cs.CL
Published: 2022-04-13
arXiv: 2204.06487v3
Abstract:
Multilingual pre-trained language models (PLMs) have demonstrated impressive performance on several downstream tasks for both high-resourced and low-resourced languages. However, there is still a large performance drop for languages unseen during pre-training, especially African languages. One of the most effective approaches to adapt to a new language is \textit{language adaptive fine-tuning} (LAFT) – fine-tuning a multilingual PLM on monolingual texts of a language using the pre-training objective. However, adapting to a target language individually takes a large disk space and limits the cross-lingual transfer abilities of the resulting models because they have been specialized for a single language. In this paper, we perform \textit{multilingual adaptive fine-tuning} on 17 most-resourced African languages and three other high-resource languages widely spoken on the African continent to encourage cross-lingual transfer learning. To further specialize the multilingual PLM, we removed vocabulary tokens from the embedding layer that corresponds to non-African writing scripts before MAFT, thus reducing the model size by around 50%. Our evaluation on two multilingual PLMs (AfriBERTa and XLM-R) and three NLP tasks (NER, news topic classification, and sentiment classification) shows that our approach is competitive to applying LAFT on individual languages while requiring significantly less disk space. Additionally, we show that our adapted PLM also improves the zero-shot cross-lingual transfer abilities of parameter efficient fine-tuning methods.
50. Looking for Clues of Language in Multilingual BERT to Improve Cross-lingual Generalization
Authors: Chi-Liang Liu, Tsung-Yuan Hsu, Yung-Sung Chuang, Chung-Yi Li, Hung-yi Lee
Categories: cs.CL, cs.AI
Published: 2020-10-20
arXiv: 2010.10041v4
Abstract:
Token embeddings in multilingual BERT (m-BERT) contain both language and semantic information. We find that the representation of a language can be obtained by simply averaging the embeddings of the tokens of the language. Given this language representation, we control the output languages of multilingual BERT by manipulating the token embeddings, thus achieving unsupervised token translation. We further propose a computationally cheap but effective approach to improve the cross-lingual ability of m-BERT based on this observation.