My research is driven by a commitment to addressing pressing societal challenges through innovative applications of artificial intelligence (AI) and machine learning (ML). I work in applied AI and data science, developing data-driven and machine-learning methods that are applied to challenges in gender equity, digital inclusion, and education.
Intensity-Aware Transformer Models for Forecasting Political Violence Events
Over the past five years, conflict monitoring and analysis has become one focus aera of my research, driven both by personal motivation and by the persistence of global conflicts. The evolution of conflict intensity—measured through the number of events and fatalities over time—poses a challenging time-series problem with significant societal implications. Reliable analysis and forecasting of such trends can support humanitarian organizations and inform prevention and response strategies.
A key research question concerns the extent to which conflict dynamics exhibit temporal regularities or seasonal patterns, despite being influenced by complex political and economic factors. This raises further questions about predictability and the suitability of advanced machine-learning models, particularly transformer-based architectures that have shown versatility across domains. In this work, I investigate whether and how such models can be applied to conflict event forecasting, and what insights they can offer for anticipatory decision-making. This paper presents the hypotheses, methodology, and findings that address these questions.
STEM Education in the Age of Generative AI
As generative AI continues to advance, higher education institutions must reconsider and realign teaching and learning practices to reflect how new generations of students acquire knowledge and skills. The increasing capability of AI systems to perform tasks such as coding raises fundamental pedagogical questions about the continued role of foundational learning levels, including the lower tiers of Bloom’s taxonomy (e.g., remembering and understanding).
This evolving landscape calls for sustained research and systematic evaluation of how generative AI reshapes learning objectives, curriculum design, and assessment in STEM education. My work examines these challenges through empirical studies on AI-supported STEM education, holistic learning approaches, and the design of customized curricula. The findings from these studies contribute evidence-based insights into how higher education can adapt responsibly and effectively in the era of generative AI.
Natural Language Processing and Information Retrieval Techniques for Low-Resource Languages
Beyond its impact on learning and teaching, generative AI also raises important technical questions about disparities between resource-rich and low-resource language contexts. As many AI systems are developed and evaluated primarily on high-resource languages, there is a risk that advances in generative and information retrieval technologies may not generalize effectively to underrepresented languages. My on-going research addresses this challenge by developing and evaluating NLP and information retrieval methods for low-resource settings, including the application of TF–IDF weighting and BM25 algorithms for Myanmar news retrieval, grammar-based text segmentation, and content-aware named entity recognition for news summarization.