Indonesian NLP Talent Breaks New Ground at ACL 2025 with Culturally Grounded AI and Award-Winning Research

Master of Data Science Students and Researchers Make Their Mark at ACL 2025

06 September 2025 | Vienna, Austria – A team of six from the Data Science Program's Natural Language Processing (NLP) research lab at Monash University, Indonesia, presented significant work at ACL 2025, the foremost conference in NLP. Led by Associate Professor Derry Wijaya, the group included Research Assistants Lucky Susanto, Musa Wijanarko, Mohammad Rifqi Farhansyah, and Master of Data Science (MDS) students Iwan Darmawan and Fariz Akyas.

Two MDS theses supervised by Associate Professor Wijaya were accepted and presented at ACL, reflecting strong scholarly contributions. Iwan Darmawan introduced Unggah-Ungguh, a dataset capturing Javanese honorific levels, assessing whether current large language models (LLMs) can recognize and generate speech appropriate to social context. The findings showed that existing models struggle and exhibit bias towards certain honorific levels, exposing a clear research gap.

Fariz Akyas presented a multi-label dataset covering Indonesian online discourse, annotated for toxicity, polarization, and annotator demographics. His work demonstrated that jointly modeling toxicity and polarization improves model performance, and incorporating demographic context further enhances detection accuracy. This has practical implications for fostering healthier online conversations in Indonesia.

Additional Research Contributions

Monash University’s NLP research group also presented two other papers at the conference. The NusaAksara project, developed in collaboration with MBZUAI, introduced the first multimodal benchmark aimed at preserving Indonesia’s indigenous scripts. It covers eight scripts (Jawa, Bali, Sunda, Batak, Lampung, Lontara, Jawi, and Pegon) and seven languages, featuring expert-validated tasks such as image segmentation, OCR/transcription, transliteration, translation, and language identification.

Compiled from 75 books spanning 7,137 pages and evaluated across top LLMs and Visual Language Models (VLMs), the study revealed near-zero performance on these scripts. This highlights a significant gap in multilingual AI capabilities and the urgent need for more inclusive AI standards. The work on Unggah-Ungguh and NusaAksara demonstrates the lab’s focus on AI grounded in cultural context.

The fourth paper, titled "Insights into Climate Change Narratives: Emotional Alignment and Engagement Analysis on TikTok," received the Best Paper Award at ACL’s NLP for Positive Impact Workshop. The study shows how emotion-aware communication strategies can boost engagement with climate-related content on social media platforms like TikTok.

Recognition and Opportunities

Adding to the achievements, Iwan Darmawan and Mohammad Rifqi Farhansyah were honored with ACL’s Diversity & Inclusion Award, sponsored by Apple. This award provided support for their travel and accommodation to present at the conference in person.

Beyond presentations, the Monash University, Indonesia team connected with researchers from academia and industry worldwide, opening doors to future research partnerships, internships, and PhD opportunities.

About the Master of Data Science Program at Monash University, Indonesia

The Master of Data Science at Monash University, Indonesia emphasizes hands-on research led by internationally recognized scholars. Students engage deeply with Indonesia’s linguistic and societal challenges, contributing to global AI discussions while ensuring local relevance.

The research lab’s portfolio—from Unggah-Ungguh and NusaAksara’s culturally grounded AI to contextual models for toxicity and polarization detection, and award-winning climate communication studies—illustrates how focused research can deliver impact both locally and globally.

For data science professionals seeking to expand their expertise in AI with a focus on real-world applications, exploring advanced courses can provide valuable skills. Visit Complete AI Training’s latest AI courses for more information.