Kidist Amde Mekonnen

I am an ELLIS PhD researcher in the Information Retrieval Lab at the University of Amsterdam, supervised by Prof. dr. Maarten de Rijke, with Dr. Andrew Yates as my co-promoter.

My research asks how retrieval systems can move beyond selecting from a fixed index and instead learn, generate, and update representations of documents and items. I work primarily on generative retrieval and recommendation, focusing on document-level relevance optimization, continual adaptation to evolving collections, robust decoding, and scalable identifier learning.

I am particularly interested in where generative retrieval systems fail: when relevant documents are lost during decoding, when adding new information degrades previously learned knowledge, and when retrieval effectiveness drops across languages or domains. My work combines model design, learning objectives, decoding algorithms, and rigorous evaluation to build retrieval systems that are more reliable, efficient, and adaptable across large, evolving, and multilingual collections.

💡 Research Highlights

• Document-level relevance optimization: designing lightweight learning objectives that directly align generative retrieval models with document-level relevance.
• Continual generative retrieval: developing parametric-memory mechanisms that incorporate new documents while preserving previously learned knowledge.
• Decoding and robustness: studying how constrained generation and look-ahead guidance affect recall under query variation and cross-lingual shift.
• Multilingual and low-resource retrieval: developing retrieval models and benchmarks for Amharic and other underrepresented languages.
• Generative recommendation: exploring semantic and multimodal item identifiers, constrained generation, and scalable identifier learning.
• Retrieval-augmented systems: contributing to multi-turn pipelines combining sparse retrieval, query reformulation, reranking, and generation.

📢 News & Updates

• 🇬🇧 November 2026 — I will undertake an ELLIS research exchange at University College London (UCL), hosted by Prof. Emine Yilmaz.
• August 2026 — I will join Booking.com as a PhD Machine Learning Research Intern, working on scalable and generalisable ID embedding learning.
• 🇦🇺 July 2026 — I will present two papers at SIGIR 2026 in Melbourne, Australia: “A Parametric Memory Head for Continual Generative Retrieval” and “Lost in Decoding? Reproducing and Stress-Testing the Look-Ahead Prior in Generative Retrieval.”
• 🇺🇸 July 2026 — Presented our paper, “The Multilingual Curse at the Retrieval Layer: Evidence from Amharic,” at MeLLM @ ACL 2026 in San Diego, USA.
• 🏆 June 2026 — Our uva-irlab-conv system, combining learned sparse retrieval with conversational query rewriting and listwise reranking, ranked 2nd out of 38 teams in the conversational search subtask of the SemEval-2026 Multi-Turn RAG Shared Task (Task 8). [Paper]
• 🔥 May 2026 — Our paper on multilingual retrieval for Amharic was accepted at MeLLM @ ACL 2026.
• 🔥 April 2026 — Two papers were accepted at SIGIR 2026: a full paper on continual generative retrieval and a reproducibility paper on decoding robustness.
• 🇦🇹 July 2025 — Presented our paper, “Optimized Text Embedding Models and Benchmarks for Amharic Passage Retrieval,” at ACL 2025 in Vienna, Austria.
• 🇮🇹 July 2025 — Presented our paper, “Lightweight and Direct Document Relevance Optimization for Generative Information Retrieval,” at SIGIR 2025 in Padova, Italy.
• 🔥 May 2025 — Our paper, “Optimized Text Embedding Models and Benchmarks for Amharic Passage Retrieval,” was accepted for publication in Findings of ACL 2025.
• 🔥 April 2025 — Our full paper, “Lightweight and Direct Document Relevance Optimization for Generative Information Retrieval,” was accepted at SIGIR 2025.

🎓 Education and Experience

• 2023–Present, ELLIS PhD Researcher, University of Amsterdam, The Netherlands
• Apr. 2022–Oct. 2022, Data Science Research Intern, Nanovery, Newcastle upon Tyne, United Kingdom
• 2020–2023, MSc in Data Science, University of Trento, Italy
• 2019–2020, MSc in Machine Intelligence, African Institute for Mathematical Sciences / AMMI, Rwanda
• 2017–2019, Assistant Lecturer, Department of Computer Science, University of Gondar, Ethiopia
• 2013–2017, BSc in Computer Science, University of Gondar, Ethiopia