Dun Li Chan
Author directory2026
Can Linguistically Related Languages Guide LLM Translation in Low-Resource Settings?
Aishwarya Ramasethu | Rohin Garg | Niyathi Allu | Harshwardhan Fartale | Dun Li Chan
Proceedings for the Ninth Workshop on Technologies for Machine Translation of Low Resource Languages (LoResMT 2026)
Aishwarya Ramasethu | Rohin Garg | Niyathi Allu | Harshwardhan Fartale | Dun Li Chan
Proceedings for the Ninth Workshop on Technologies for Machine Translation of Low Resource Languages (LoResMT 2026)
Large Language Models (LLMs) have achieved strong performance across many downstream tasks, yet their effectiveness in extremely low-resource machine translation remains limited. Standard adaptation techniques typically rely on large-scale parallel data or extensive fine-tuning, which are infeasible for the long tail of underrepresented languages. In this work, we investigate a more constrained question: in data-scarce settings, to what extent can linguistically similar pivot languages and few-shot demonstrations provide useful guidance for on-the-fly adaptation in LLMs? We study a data-efficient experimental setup that combines linguistically related pivot languages with few-shot in-context examples, without any parameter updates, and evaluate translation behavior under controlled conditions. Our analysis shows that while pivot-based prompting can yield improvements in certain configurations, particularly in settings where the target language is less well represented in the model’s vocabulary, the gains are often modest and sensitive to few shot example construction. For closely related or better represented varieties, we observe diminishing or inconsistent gains. Broadly, our findings provide empirical guidance on how and when inference-time prompting and pivot-based examples can be used as a lightweight alternative to fine-tuning in low-resource translation settings.
Anthropogenic Regional Adaptation in Multimodal Vision-Language Model
Samuel Cahyawijaya | Peerat Limkonchotiwat | Tack Hwa Wong | Carlos Rafael Catalan | Manuel Antonio Rufino | Hitesh Laxmichand Patel | Amit Agarwal | Muhammad Reza Qorib | Vicky Feliren | Holy Lovenia | Aye Hninn Khine | Frederikus Hudi | David Anugraha | Alham Fikri Aji | Romrawin Chumpu | Viet Thanh Pham | Minghan Wang | Mohamed Fazli Mohamed Imam | Ruochen Zhang | Joseph Marvin Imperial | Khumaisa Nur’aini | Do Xuan Long | Musa Izzanardi Wijanarko | Joel Ruben Antony Moniz | Patrick Amadeus Irawan | Hanif Muhammad Zhafran | Isaiah Edri W. Flores | Salsabila Zahirah Pranida | Jun Kevin | Jostin Jerico Rosal | Patricia Nicole Monderin | Kun Kerdthaisong | Ahmad Mustafid | My Chiffon Nguyen | Natchapon Jongwiriyanurak | Siva Worajitwannakul | Haochen Li | Adrian Xuan Wei Lim | Bin Wang | Muhammad Ravi Shulthan Habibi | Lynnette Hui Xian Ng | Mithil Bangera | Yeshil Bangera | Priyaranjan Pattnayak | Dun Li Chan | Sherissa Caren Djuniwar | Cho Chan Myei Oo | Ming Shan Hee
Findings of the Association for Computational Linguistics: IJCNLP-AACL 2026
Samuel Cahyawijaya | Peerat Limkonchotiwat | Tack Hwa Wong | Carlos Rafael Catalan | Manuel Antonio Rufino | Hitesh Laxmichand Patel | Amit Agarwal | Muhammad Reza Qorib | Vicky Feliren | Holy Lovenia | Aye Hninn Khine | Frederikus Hudi | David Anugraha | Alham Fikri Aji | Romrawin Chumpu | Viet Thanh Pham | Minghan Wang | Mohamed Fazli Mohamed Imam | Ruochen Zhang | Joseph Marvin Imperial | Khumaisa Nur’aini | Do Xuan Long | Musa Izzanardi Wijanarko | Joel Ruben Antony Moniz | Patrick Amadeus Irawan | Hanif Muhammad Zhafran | Isaiah Edri W. Flores | Salsabila Zahirah Pranida | Jun Kevin | Jostin Jerico Rosal | Patricia Nicole Monderin | Kun Kerdthaisong | Ahmad Mustafid | My Chiffon Nguyen | Natchapon Jongwiriyanurak | Siva Worajitwannakul | Haochen Li | Adrian Xuan Wei Lim | Bin Wang | Muhammad Ravi Shulthan Habibi | Lynnette Hui Xian Ng | Mithil Bangera | Yeshil Bangera | Priyaranjan Pattnayak | Dun Li Chan | Sherissa Caren Djuniwar | Cho Chan Myei Oo | Ming Shan Hee
Findings of the Association for Computational Linguistics: IJCNLP-AACL 2026
While the field of vision-language (VL) has achieved remarkable success in integrating visual and textual information across multiple languages and domains, there is no dedicated framework for assessing human-centric alignment in vision-language systems.
We offer two contributions to address this gap. First, we introduce Anthropogenic Regional Adaptation: a regional-adaptation framework that aims to optimize model relevance to specific regional contexts while ensuring the retention of global generalization capabilities. Second, we present a simple, but effective adaptation method named Geographical-generalization-made-easy (GG-EZ), which utilizes regional data filtering and model merging.
Through comprehensive ablations on 3 VL architectures: large vision-language models (LVLM), text-to-image diffusion models, and vision-language embedding models, and a case study in Southeast Asia (SEA) regional adaptation, we demonstrate the importance of Anthropogenic Regional Adaptation and GG-EZ, showing 5-15% gains in cultural relevance metrics across SEA while maintaining over 98% of global performance and even occasionally surpassing it.
Our findings establish Anthropogenic Regional Alignment as a foundational paradigm towards applicability of multimodal VL architectures in diverse regions and demonstrate a simple-yet-effective baseline method that optimizes regional value alignment while preserving global generalization.
We offer two contributions to address this gap. First, we introduce Anthropogenic Regional Adaptation: a regional-adaptation framework that aims to optimize model relevance to specific regional contexts while ensuring the retention of global generalization capabilities. Second, we present a simple, but effective adaptation method named Geographical-generalization-made-easy (GG-EZ), which utilizes regional data filtering and model merging.
Through comprehensive ablations on 3 VL architectures: large vision-language models (LVLM), text-to-image diffusion models, and vision-language embedding models, and a case study in Southeast Asia (SEA) regional adaptation, we demonstrate the importance of Anthropogenic Regional Adaptation and GG-EZ, showing 5-15% gains in cultural relevance metrics across SEA while maintaining over 98% of global performance and even occasionally surpassing it.
Our findings establish Anthropogenic Regional Alignment as a foundational paradigm towards applicability of multimodal VL architectures in diverse regions and demonstrate a simple-yet-effective baseline method that optimizes regional value alignment while preserving global generalization.
Search
Co-authors
- Amit Agarwal 1
- Alham Fikri Aji 1
- Niyathi Allu 1
- David Anugraha 1
- Mithil Bangera 1
- Yeshil Bangera 1
- Samuel Cahyawijaya 1
- Carlos Rafael Catalan 1
- Romrawin Chumpu 1
- Sherissa Caren Djuniwar 1
- Harshwardhan Fartale 1
- Vicky Feliren 1
- Isaiah Edri W. Flores 1
- Rohin Garg 1
- Muhammad Ravi Shulthan Habibi 1
- Ming Shan Hee 1
- Frederikus Hudi 1
- Mohamed Fazli Mohamed Imam 1
- Joseph Marvin Imperial 1
- Patrick Amadeus Irawan 1
- Natchapon Jongwiriyanurak 1
- Kun Kerdthaisong 1
- Jun Kevin 1
- Aye Hninn Khine 1
- Haochen Li 1
- Adrian Xuan Wei Lim 1
- Peerat Limkonchotiwat 1
- Do Xuan Long 1
- Holy Lovenia 1
- Patricia Nicole Monderin 1
- Joel Ruben Antony Moniz 1
- Ahmad Mustafid 1
- Lynnette Hui Xian Ng 1
- My Chiffon Nguyen 1
- Khumaisa Nur’aini 1
- Cho Chan Myei Oo 1
- Hitesh Laxmichand Patel 1
- Priyaranjan Pattnayak 1
- Viet Thanh Pham 1
- Salsabila Zahirah Pranida 1
- Muhammad Reza Qorib 1
- Aishwarya Ramasethu 1
- Jostin Jerico Rosal 1
- Manuel Antonio Rufino 1
- Bin Wang 1
- Minghan Wang 1
- Musa Izzanardi Wijanarko 1
- Tack Hwa Wong 1
- Siva Worajitwannakul 1
- Hanif Muhammad Zhafran 1
- Ruochen Zhang 1