Published in SynthIR Workshop, ACM SIGIR, 2026
A study of the trade-off between retrieval effectiveness and faithfulness when large language models generate or augment RDF dataset metadata.
Recommended citation: Riccardo Terrenzi and Serkan Ayvaz. (2026). "Faithful or Findable? Evaluating LLM-Generated Metadata for RDF Dataset Search." SynthIR Workshop at ACM SIGIR 2026.
Download Paper
Published in Database and Expert Systems Applications (DEXA), 2026
PIPER uses table profiles and LLM-generated pseudoqueries for content-driven dataset retrieval when metadata is sparse or low quality.
Recommended citation: Riccardo Terrenzi, Matteo Falconi, Serkan Ayvaz, and Pierluigi Plebani. (2026). "PIPER: Content-Based Table Search via Profiling and LLM-Generated Pseudoqueries." DEXA 2026.
Download Paper
Published in GenAIK-NORA Workshop, IJCAI-ECAI, 2026
An investigation of citation faithfulness in Agentic GraphRAG showing that answers can depend on uncited traversal context and surrounding graph structure.
Recommended citation: Riccardo Terrenzi, Maximilian von Zastrow, and Serkan Ayvaz. (2026). "Why Neighborhoods Matter: Traversal Context and Provenance in Agentic GraphRAG." GenAIK-NORA Workshop at IJCAI-ECAI 2026.
Download Paper
Published in SAML Workshop, IEEE International Conference on Software Architecture, 2026
A bounded and auditable architecture for dataset search combining lexical and dense retrieval with an LLM agent that plans, evaluates, and reranks results.
Recommended citation: Riccardo Terrenzi, Phongsakon Mark Konrad, Tim Lukas Adam, and Serkan Ayvaz. (2026). "A Reference Architecture for Agentic Hybrid Retrieval in Dataset Search." 2026 IEEE 23rd International Conference on Software Architecture Companion, 542-548.
Download Paper
Published in KDA-AI Workshop, IEEE International Conference on Software Architecture, 2026
CAKE evaluates how well large language models understand cloud-native software architecture.
Recommended citation: Tim Lukas Adam, Phongsakon Mark Konrad, Riccardo Terrenzi, Florian Girardo Lukas, Rahime Yilmaz, Krzysztof Sierszecki, and Serkan Ayvaz. (2026). "CAKE: Cloud Architecture Knowledge Evaluation of Large Language Models." 2026 IEEE 23rd International Conference on Software Architecture Companion, 361-368.
Download Paper
Published in SAGAI Workshop, IEEE International Conference on Software Architecture, 2026
AI coding agents make implicit architectural choices through framework selection, infrastructure scaffolding, and integration decisions. This work introduces the idea of vibe architecting and proposes practices for governing prompt-driven architectural decisions.
Recommended citation: Phongsakon Mark Konrad, Tim Lukas Adam, Riccardo Terrenzi, and Serkan Ayvaz. (2026). "Architecture Without Architects: How AI Coding Agents Shape Software Architecture." 2026 IEEE 23rd International Conference on Software Architecture Companion, 463-468.
Download Paper
Published in arXiv preprint, 2026
A proposal for calibrating AI verification requirements to deployment risk rather than treating model transparency as a binary prerequisite.
Recommended citation: Phongsakon Mark Konrad, Tim Lukas Adam, Ane Cathrine Holst Merrild, Riccardo Terrenzi, Rebecca De Rosa, Toygar Tanyel, and Serkan Ayvaz. (2026). "The Open-Box Fallacy: Why AI Deployment Needs a Calibrated Verification Regime." arXiv preprint arXiv:2605.10601.
Download Paper