Emotion in LLMs
Emotion understanding, appraisal-style reasoning, and response generation for emotionally sensitive applications.
Rumana Ferdous Munne
Postdoctoral Researcher
Language and Information Research Team
Artificial Intelligence Research Center (AIRC), AIST
Tokyo, Japan
Email: rf.munne@aist.go.jp
I am an NLP researcher specializing in LLM reasoning, document-level information extraction, entity alignment, and knowledge representation, with current research spanning scientific and biomedical NLP, ontology-guided reasoning, low-resource information extraction, and emotion-aware text generation for LLMs. My work studies how language models and structured knowledge can infer implicit meaning from complex, domain-specific documents and transform it into structured, interpretable, and useful representations.
My background combines 14+ years of AI research and software engineering experience across AIST, RIKEN, the National Institute of Informatics, and Samsung Electronics.
I welcome inquiries, collaborations, and research opportunities.
Experience
-
Apr 2026 – Present
Postdoctoral Researcher
Artificial Intelligence Research Center (AIRC), AIST. Language and Information Research Team, P.I.: Prof. Hiroya Takamura.
- Study emotion understanding and emotion-aware text generation for LLMs, focusing on appraisal-style reasoning and response generation for emotionally sensitive applications.
-
Sep 2023 – Mar 2026
Postdoctoral Researcher
RIKEN Center for Advanced Intelligence Project, Japan. Knowledge Acquisition Team, P.I.: Prof. Yuji Matsumoto.
- Researched NLP methodologies for low-resource scientific and biomedical information extraction, focusing on implicit knowledge, ontology-guided verification, and domain-specific constraints.
-
Apr 2022 – Jul 2023
Special Researcher
Tokyo Institute of Technology, Japan
- Led collaborative research between the National Institute of Informatics and Tokyo Institute of Technology.
-
Oct 2016 – Mar 2023
Research Assistant
National Institute of Informatics, Japan
- Designed algorithms for knowledge graph alignment and completion while completing Ph.D. research in Informatics.
-
Dec 2011 – Aug 2016
Software Engineer / Senior Software Engineer
Samsung R&D Institute Bangladesh and Samsung Electronics, South Korea
- Developed Android, Tizen, wearable device, and Galaxy smartphone platform features.
Research Areas
LLMs for structured knowledge reasoning
Entity profile generation, LLM-based reasoning, and refinement methods for knowledge graph and information extraction tasks.
Biomedical information extraction
Ontology-driven concept recognition and low-resource annotation from biomedical and scientific literature.
AI for scientific discovery
NLP systems that turn scientific text into structured, reusable knowledge for search, discovery, and downstream reasoning.
Selected Work
Entity Profile Generation and Reasoning with LLMs for Entity Alignment
R. F. Munne, M. M. Rahman, Y. Matsumoto.
Dissecting GraphRAG for Factoid Question Answering
N. Nishida, R. F. Munne, S. Liu, N. Tokunaga, Y. Yamagata, F. Cheng, K. Kozaki, Y. Matsumoto.
Zero-Shot Entailment Learning for Biomedical Annotation
R. F. Munne, N. Nishida, S. Liu, N. Tokunaga, Y. Yamagata, K. Kozaki, Y. Matsumoto.
Lookalike Audience Expansion: A Graph-Based Model with LLMs
R. F. Munne, M. M. Rahman, Y. Matsumoto.
Education
Ph.D. in Informatics, National Institute of Informatics, Japan, 2023
M.Sc. in Informatics, National Institute of Informatics, Japan, 2020
M.Sc. in CSE, University of Dhaka, Bangladesh, 2014
B.Sc. in CSE, University of Dhaka, Bangladesh, 2011
Honors
Monbukagakusho (MEXT) Scholarship, 2016
Young Leadership Award, 2015
Samsung R&D Iconic Team Member, 2013
University of Dhaka Outstanding Result Scholarship, 2011