Research

Research Profile

Poli Nemkova is an AI researcher working at the intersection of multilingual natural language processing, robust and responsible AI, and AI for Social Good. Her research focuses on developing and evaluating AI systems for decision support in complex social systems with applications to conflict forecasting, human rights monitoring.  She combines methodological contributions in NLP and machine learning with large-scale empirical studies, including dataset construction in low-resource languages, and have published at leading AI venues (AAAI, IJCAI, LREC, EACL, NeurIPS).

Research Agenda

My research focuses on developing reliable, multilingual, and socially responsible AI systems for decision support in high-stakes domains. I aim to bridge advances in machine learning and natural language processing with real-world deployment constraints in humanitarian, policy, and societal contexts. My research agenda is organized around four interrelated directions:

• Agentic and uncertainty-aware AI systems for decision support. Agentic architectures integrating uncertainty estimation, metacognitive reasoning, and multi-trajectory inference to improve robustness and interpretability in forecasting and decision-making.

• Multilingual and low-resource NLP for conflict forecasting and social good Methods for extracting signals from unstructured multilingual text streams; focus on low-resource languages and domain shift in conflict, human rights, and misinformation monitoring.

• Evaluation, robustness, and responsible deployment of foundation models. Systematic evaluation frameworks for LLMs and generative models, including bias, stability, and failure modes in high-stakes social applications.

Selected Publications

LIME-LLM: Probing Models with Fluent Counterfactuals, Not Broken Text
G. Mihaila, S. O. Polat, P. Nemkova, H. Sharma, N. A. Urs, M. V. Albert
Under submission

Cross-Lingual Stability and Bias in Instruction-Tuned Language Models for Humanitarian NLP
P. Nemkova, A. Adhikari, M. Pearson, V. Krishna Sadu, A. V. Mark
LREC 2026 Main Track

Agentic Multilingual NLP for Conflict Forecasting from Open-Source Text Streams
P. A. Nemkova, M. V. Albert
WiML @ NeurIPS 2025

NLP for Social Good: A Survey of Challenges, Opportunities, and Responsible Deployment
A. Karamolegkou, A. Borah, E. Cho, S. R. Choudhury, … P. Nemkova, … D. Dementieva (2025)
EACL 2026 · Main Track

Comparing LLM Text Annotation Skills: A Study on Human Rights Violations in Social Media Data
P. A. Nemkova, S. Ubani, M. V. Albert
AAAI 2025

Detecting Human Rights Violations on Social Media Using Machine Learning
P. A. Nemkova, S. Polat, S. Ubani, K. Nguen, R. Nielsen
AAAI 2025

Do Large Language Models Know Conflict? Investigating Parametric vs. Non-Parametric Knowledge of LLMs for Conflict Forecasting
P. A. Nemkova, S. C. Lingareddy, S. Ray Choudhury, M. V. Albert
2025

Towards Automated Situation Awareness: A RAG-Based Framework for Peacebuilding Reports
P. A. Nemkova, S. O. Polat, R. I. Jahan, S. R. Choudhury, S. Lee, S. Sarkar, M. V. Albert (2025)
Under submission · 2025

Synthetic Adaptive Guided Embeddings (SAGE): A Novel Knowledge Distillation Method
S. O. Polat, P. A. Nemkova, M. V. Albert
Under submission · 2025

Improved Bitcoin Price Prediction based on COVID-19 data
P. Nemkova, R. Moreira
IJCAI-PRICAI 2021

Conference Reviewing

• Program Committee: AIES 2025

• Reviewer: ICECET 2025, ICLR 2025, ACL ARR 2025, EMNLP 2023, ACM 2021, LREC 2026

Grants & External Funding

• Spring 2026 — Grant for supporting student researcher assistants in Nemkova's Lab 42 research good.

• May 2025 — Routledge (Taylor & Francis Group): Book contract signed (Collaboration: CRRIC)

• Nov 2023 — Digital Citizen Contribution Program (Canadian Heritage): “Truth Guard” (Collaboration: CRRIC)

• Aug 2025 — Digital Citizen Contribution Program (Canadian Heritage): “M³ LLM Project” (Collaboration: CRRIC) — Under review