LLMs for Low Resource Languages

Cite the tutorial

If you use material from this tutorial, please cite:

Alam, Firoj and Shammur Absar Chowdhury. Multilingual and Multimodal LLMs in the Wild: Building for Low-Resource Languages. LREC 2026 Tutorial. https://mm-llms-in-the-wild.github.io

Bib Entry 

@misc{alam_chowdhury_2026_mm_llms_wild,
  title        = {Multilingual and Multimodal LLMs in the Wild: Building for Low-Resource Languages},
  author       = {Alam, Firoj and Chowdhury, Shammur Absar},
  year         = {2026},
  howpublished = {\url{https://mm-llms-in-the-wild.github.io}},
  note         = {LREC 2026 Tutorial: https://lrec2026.info/lrec2026-tutorials/}
}

Suggested reading

1 · Introduction — low-resource languages


2 · LLM foundations

  • Brown et al. (2020) — Language Models are Few-Shot Learners (GPT-3). NeurIPS 2020. NeurIPS Proceedings · arXiv:2005.14165
  • Zhao et al. (2023) — A Survey of Large Language Models. arXiv. arXiv:2303.18223
  • Bubeck et al. (2023) — Sparks of Artificial General Intelligence: Early Experiments with GPT-4. arXiv. arXiv:2303.12712
  • Liang et al. (2022) — Holistic Evaluation of Language Models (HELM). TMLR 2023. arXiv:2211.09110


3 · Multilingual / regional LLMs

  • Wei et al. (2023) — PolyLM: An Open Source Polyglot Large Language Model. arXiv. arXiv:2307.06018
  • Nguyen, Xuan-Phi et al. (2023) — SeaLLMs: Large Language Models for Southeast Asia. arXiv. arXiv:2312.00738


4 · Pre-training data & multilingual resources


5 · Visual / chart / document benchmarks

  • Antol et al. (2015) — VQA: Visual Question Answering. ICCV 2015. CVF Open Access · arXiv:1505.00468
  • Mathew et al. (2021) — DocVQA: A Dataset for VQA on Document Images. WACV 2021. CVF Open Access · arXiv:2007.00398
  • Yue et al. (2024) — MMMU: A Massive Multi-discipline Multimodal Understanding and Reasoning Benchmark for Expert AGI. CVPR 2024. CVF Open Access · arXiv:2311.16502
  • Yue et al. (2024) — MMMU-Pro: A More Robust Multi-discipline Multimodal Understanding Benchmark. arXiv. arXiv:2409.02813
  • Liu et al. (2023) — MMBench: Is Your Multi-modal Model an All-around Player? arXiv. arXiv:2307.06281 · https://github.com/open-compass/MMBench
  • Masry et al. (2022) — ChartQA: A Benchmark for Question Answering about Charts with Visual and Logical Reasoning. ACL 2022 Findings. ACL Anthology 2022.findings-acl.177 · arXiv:2203.10244
  • Masry et al. (2025) — ChartQAPro: A More Diverse and Challenging Benchmark for Chart Question Answering. ACL 2025 Findings. arXiv:2504.05506
  • Wang et al. (2024) — CharXiv: Charting Gaps in Realistic Chart Understanding in Multimodal LLMs. NeurIPS 2024 D&B. OpenReview cy8mq7QYae · arXiv:2406.18521
  • Kantharaj et al. (2022) — OpenCQA: Open-ended Question Answering with Charts. EMNLP 2022. ACL Anthology 2022.emnlp-main.811 · arXiv:2210.06628
  • Mathew et al. (2022) — InfographicVQA. WACV 2022. CVF Open Access · arXiv:2104.12756
  • Hsiao et al. (2022) — ScreenQA: Large-Scale Question-Answer Pairs over Mobile App Screenshots. arXiv. arXiv:2209.08199
  • Tanaka et al. (2023) — SlideVQA: A Dataset for Document Visual Question Answering on Multiple Images. AAAI 2023. AAAI Proceedings · arXiv:2301.04883
  • Kantharaj et al. (2022) — Chart-to-Text: A Large-Scale Benchmark for Chart Summarization. ACL 2022. ACL Anthology 2022.acl-long.277
  • Tang et al. (2023) — VisText: A Benchmark for Semantically Rich Chart Captioning. ACL 2023. arXiv:2307.05356
  • Rahman et al. (2025) — Text2Vis: A Challenging and Diverse Benchmark for Generating Multimodal Visualizations from Text. EMNLP 2025. arXiv:2507.19969
  • Xie, Tianbao et al. (2024) — OSWorld: Benchmarking Multimodal Agents for Open-Ended Tasks in Real Computer Environments. NeurIPS 2024. arXiv:2404.07972
  • Kartha et al. (2026) — DashboardQA: Benchmarking Multimodal Agents for QA on Interactive Dashboards. EACL 2026 Findings. (https://aclanthology.org/2026.findings-eacl.177/)
  • Fu et al. (2024) — Video-MME: The First-Ever Comprehensive Evaluation Benchmark of Multi-modal LLMs in Video Analysis. CVPR 2025. arXiv:2405.21075
  • Li, Kunchang et al. (2024) — MVBench: A Comprehensive Multi-modal Video Understanding Benchmark. CVPR 2024. arXiv:2311.17005


6 · Multilingual / low-resource challenges

  • Laskar, Islam, Nayeem, Bhuiyan, Rahman, Joty, Hoque, Huang (2026) — Lost in Translation: Do LVLM Judges Generalize Across Languages? (MM-JudgeBench). ACL 2026 (accepted).


7 · Multimodal LLM architectures & examples


8 · Speech tokenisation & audio codecs

  • Zeghidour et al. (2022) — SoundStream: An End-to-End Neural Audio Codec. IEEE/ACM TASLP 2022. arXiv:2107.03312
  • Défossez, Alexandre et al. (2022) — High Fidelity Neural Audio Compression (EnCodec). arXiv. arXiv:2210.13438
  • Hsu et al. (2021) — HuBERT: Self-Supervised Speech Representation Learning by Masked Prediction of Hidden Units. IEEE/ACM TASLP 2021. arXiv:2106.07447
  • Baevski et al. (2020) — wav2vec 2.0: A Framework for Self-Supervised Learning of Speech Representations. NeurIPS 2020. NeurIPS Proceedings · arXiv:2006.11477
  • Chung et al. (2021) — w2v-BERT: Combining Contrastive Learning and Masked Language Modeling for Self-Supervised Speech Pre-Training. ASRU 2021. arXiv:2108.06209
  • Zhang, Xin et al. (2024) — SpeechTokenizer: Unified Speech Tokenizer for Speech Language Models. ICLR 2024. arXiv:2308.16692


9 · Multilingual speech models


10 · Modality generators (image / audio / video)


11 · MM-LLM surveys

  • Zhang, Duzhen et al. (2024) — MM-LLMs: Recent Advances in Multimodal Large Language Models. arXiv. arXiv:2401.13601
  • Xie, Junlin et al. (2024) — Large Multimodal Agents: A Survey. arXiv. arXiv:2402.15116
  • Wu, Jiayang et al. (2023) — Multimodal Large Language Models: A Survey. IEEE BigData 2023. arXiv:2311.13165
  • Yin, Shukang et al. (2023) — A Survey on Multimodal Large Language Models. arXiv. arXiv:2306.13549


12 · Prompting & benchmarking tools


13 · Cultural bias


14 · Misinformation

  • Chen, Canyu & Shu, Kai (2023) — Combating Misinformation in the Age of LLMs: Opportunities and Challenges. AI Magazine 2024. arXiv:2311.05656
  • Chen, Canyu & Shu, Kai (2024) — Can LLM-Generated Misinformation Be Detected? ICLR 2024. arXiv:2309.13788


15 · Hallucination


16 · Jailbreaking / red-teaming

  • Zou et al. (2023) — Universal and Transferable Adversarial Attacks on Aligned Language Models (AdvBench). arXiv. arXiv:2307.15043
  • Yong et al. (2023) — Low-Resource Languages Jailbreak GPT-4. arXiv. arXiv:2310.02446


17 · Compute

  • Luccioni, Viguier, Ligozat (2023) — Estimating the Carbon Footprint of BLOOM, a 176B Parameter Language Model. JMLR 2023. arXiv:2211.02001


Tools

Datasets & Benchmarks

Pre-training corpora

Visual / chart / document benchmarks

Multilingual & culture-aware benchmarks

Speech & audio benchmarks

Evaluation suites