# Mehmet Onur Keskin - Comprehensive Research & Academic Profile > AI researcher working on how autonomous agents can perceive, model, and adapt to human behavior in order to make better joint decisions. Research sits at the intersection of multi-agent systems, affective computing, and adaptive information retrieval, with a focus on building agents that operate reliably under preference uncertainty and real-time constraints. - **Website**: https://monurkeskin.com - **Google Scholar**: https://scholar.google.com/citations?user=jGrLNAEAAAAJ - **LinkedIn**: https://www.linkedin.com/in/monurkeskin - **X (Twitter)**: https://x.com/MehmetOnurKesk1 - **ORCID**: https://orcid.org/0000-0002-8025-4304 - **Semantic Scholar**: https://www.semanticscholar.org/author/2086876375 - **ResearchGate**: https://www.researchgate.net/profile/Mehmet-Onur-Keskin-2 - **Email**: onur.keskin@ozu.edu.tr - **Location**: Amsterdam, the Netherlands --- ## 1. Biography & Academic Trajectory Mehmet Onur Keskin earned an M.Sc. in Artificial Intelligence *(summa cum laude)* under the supervision of Dr. Reyhan Aydogan at the Interactive Intelligent Systems Lab, Ozyegin University. His thesis addressed a core challenge in human-agent negotiation: how autonomous agents can interpret human preferences, verbal arguments, and nonverbal affective signals in real time to adapt their negotiation strategies. His academic trajectory includes multiple international research appointments: - **TU Delft (Interactive Intelligence Group, Prof. Catholijn Jonker)**: Investigated multimodal human-agent negotiation frameworks through controlled laboratory experiments involving 60+ participants, analyzing joint reasoning over speech, facial expressions, and offer trajectories. - **University of Cambridge (Affective Intelligence & Robotics Lab, Prof. Hatice Gunes)**: Applied continual learning methodologies to personalized negotiation agents. Translated real-time arousal and valence signals from user facial expressions into dynamic decision functions that prevent catastrophic forgetting across repeated human-robot interactions. - **Brown University (E-GLAMOR Group, Prof. Amy Greenwald)**: Developed autonomous supply chain agents. Designed game-theoretic combinatorial dispatching algorithms aligning short-term market transactions with long-term trust and profitability objectives (Runner-up in the IJCAI 2021 Supply Chain Management League). - **Oregon State University (CoRIS Institute, NSF REU Scholar, Prof. Yigit Menguc)**: Formulated optimization algorithms for robotic 3D printing, exploring decision-making and planning within physical, embodied systems. In industry, he co-founded Kanzy Financial Technologies (London, UK) and served as CTO, leading the architecture of Finguru, an AI-assisted brokerage platform. His work centered on multi-agent orchestration, persistent memory architectures for user preference modelling, and risk-aware decision-making within strict financial compliance frameworks (supported by NVIDIA Inception and Google for Startups AI). --- ## 2. Research Dimensions ### Dimension A: Multi-Agent Systems & Automated Negotiation - Bidding strategies under incomplete preference information. - Argumentation-based negotiation extracting human speech acts in chat environments. - Decentralized multi-agent path finding (MAPF) through bilateral negotiation protocols. ### Dimension B: Affective Computing & Human-Robot Interaction - Continual learning on streaming affective signals (valence and arousal). - Embodied agent interactions on physical humanoid robots (Nao, Pepper, QT). - The influence of robot embodiment and visual appearance on negotiation dynamics. ### Dimension C: Adaptive Information Retrieval & RAG - Joint modality (textual vs. multimodal) and architecture (dense vs. late-interaction) query routing. - Empirical benchmarking of modern RAG pipelines under latency and accuracy constraints. --- ## 3. Publications Bibliography 1. **RetrievalRouter: Joint Modality and Architecture Selection for Document Retrieval** (EMNLP 2026) - *Authors*: Emre Kuru, Mehmet Onur Keskin, Reza Farahbakhsh, Noel Crespi - *Venue*: Conference on Empirical Methods in Natural Language Processing (EMNLP 2026) 2. **Personalized Affective Perception via Continual Learning in Human-Agent Negotiation** (2026) - *Authors*: Mehmet Onur Keskin, Berk Buzcu, Nikhil Churamani, Reyhan Aydogan, Hatice Gunes, Catholijn M. Jonker - *Venue*: IEEE Transactions on Affective Computing (Under Review) 3. **Evaluating Modern RAG: Textual, Multimodal, Dense, and Late Interaction Pipelines** (2025) - *Authors*: Emre Kuru, Mehmet Onur Keskin - *Venue*: 48th International ACM SIGIR Conference on Research and Development in Information Retrieval, Workshop on Information Retrieval's Role in RAG Systems - *Preprint*: arXiv:2608.23176 4. **An Adaptive Emotion-Aware Strategy for Human-Agent Negotiation: Insights from Real-World Human-Robot Experiments** (2025) - *Authors*: Mehmet Onur Keskin, Umut Çakan, Reyhan Aydogan - *Venue*: 25th ACM International Conference on Intelligent Virtual Agents (IVA 2025) - *DOI*: 10.1145/3717511.3747087 5. **Taking into Account Opponent's Arguments in Human-Agent Negotiations** (2025) - *Authors*: Anıl Doğru, Mehmet Onur Keskin, Reyhan Aydogan - *Venue*: ACM Transactions on Interactive Intelligent Systems (ACM TiiS 2025) - *DOI*: 10.1145/3691643 6. **NEGOTIATOR: A Comprehensive Framework for Human-Agent Negotiation Integrating Preferences, Interaction, and Emotion** (2024) - *Authors*: Mehmet Onur Keskin, Berk Buzcu, Berkecan Koçyiğit, Umut Çakan, Anıl Doğru, Reyhan Aydogan - *Venue*: 33rd International Joint Conference on Artificial Intelligence (IJCAI 2024) - *DOI*: 10.24963/ijcai.2024/1012 7. **NegoLog: An Integrated Python-based Automated Negotiation Framework with Enhanced Assessment Components** (2024) - *Authors*: Berk Buzcu, Berkecan Koçyiğit, Mehmet Onur Keskin, Reyhan Aydogan - *Venue*: 33rd International Joint Conference on Artificial Intelligence (IJCAI 2024) 8. **You Look Nice, but I Am Here to Negotiate: The Influence of Robot Appearance on Negotiation Dynamics** (2024) - *Authors*: Selen Akay, Mehmet Onur Keskin, Ayşe Doğan, Reyhan Aydogan, Hatice Gunes, Catholijn Jonker, Junko Kanero - *Venue*: ACM/IEEE International Conference on Human-Robot Interaction (HRI 2024) - *DOI*: 10.5555/3610978.3640700 9. **Decentralized Multi-Agent Path Finding Framework and Strategies Based on Automated Negotiation** (2023) - *Authors*: Cihan Eran, Mehmet Onur Keskin, Reyhan Aydogan - *Venue*: Autonomous Agents and Multi-Agent Systems (JAAMAS 2023) - *DOI*: 10.1007/s10458-023-09605-7 10. **Effects of Agent's Embodiment in Human-Agent Negotiations** (2023) - *Authors*: Mehmet Onur Keskin, Umut Çakan, Reyhan Aydogan - *Venue*: 23rd ACM International Conference on Intelligent Virtual Agents (IVA 2023) - *DOI*: 10.1145/3577190.3614138 11. **Conflict-Based Negotiation Strategy for Human-Agent Negotiation** (2023) - *Authors*: Anıl Doğru, Furkan Cantürk, Mehmet Onur Keskin, Reyhan Aydogan - *Venue*: IEEE Transactions on Human-Machine Systems (IEEE THMS 2023) - *DOI*: 10.1109/THMS.2023.3276632 12. **Time Series Predictive Models for Opponent Behavior Modeling in Bilateral Negotiations** (2022) - *Authors*: Mehmet Onur Keskin, Berk Buzcu, Reyhan Aydogan - *Venue*: Principles and Practice of Multi-Agent Systems (PRIMA 2022) - *DOI*: 10.1007/978-3-031-21203-1_29 13. **Would You Imagine Yourself Negotiating With a Robot, Jennifer? Why Not?** (2022) - *Authors*: Mehmet Onur Keskin, Berk Buzcu, Reyhan Aydogan - *Venue*: 31st IEEE International Conference on Robot and Human Interactive Communication (RO-MAN 2022) - *DOI*: 10.1109/RO-MAN53752.2022.9900609 14. **A Decentralized Token-Based Negotiation Approach for Multi-Agent Path Finding** (2021) - *Authors*: Mehmet Onur Keskin, Cihan Eran, Reyhan Aydogan - *Venue*: 18th European Conference on Multi-Agent Systems (EUMAS 2021) - *Award*: Second Best Paper Award 15. **Solver Agent: Towards Emotional and Opponent-Aware Agent for Human-Robot Negotiation** (2021) - *Authors*: Mehmet Onur Keskin, Reyhan Aydogan - *Venue*: 20th International Conference on Autonomous Agents and Multiagent Systems (AAMAS 2021) --- ## 4. Honors, Awards & Grants - **Second Best Paper Award** - EUMAS 2021 - **1st Place Winner** - IBM Watson Human Multi-Agent Immersive Negotiation Challenge at IJCAI 2020 - **Runner-up** - Supply Chain Management League at IJCAI 2021 (SCML 2021) - **IEEE CS Richard Merwin Student Scholarship** - **Young Researcher Participant** - Heidelberg Laureate Forum (HLF) - **Best Research Assistant Award** - Ozyegin University - **NSF REU Scholar** - Oregon State University