Simone Papicchio

Argomento di ricerca

Tecniche di contrastive learning per la rappresentazione di dati tabulari.

Tutori

• Luca Cagliero
• Paolo Papotti

Presentazione della ricerca

Poster

• I was invited to present our work, "QATCH: Benchmarking SQL-Centric Tasks with Table Representation Learning Models on Your Data," at the Gray Systems Lab at Microsoft. This work was published at NeurIPS, a CORE A*-ranked conference recognized as one of the most prestigious venues in machine learning. (2023)
• I had the opportunity to spend three months as a Research Scientist Intern at Megagon Labs in California, where I developed a framework for automatically evaluating retrieval-augmented generation (RAG) systems across different input modalities. This work culminated in a paper published at a CORE A*-ranked conference, "CypherBench: Towards Precise Retrieval over Full-Scale Modern Knowledge Graphs in the LLM Era." (2024)
• I was honored to be selected as one of only 36 participants for the invite-only workshop "LLMs Meet Data Processing" at the University of California, Berkeley. The workshop brought together researchers from leading institutions, including Berkeley and Stanford, as well as representatives from prominent companies such as Google and Databricks. (2025)
• I was involved for 1 year with three other PhD students in the research project "AI4TA: Content annotation and summarization," funded by the Intesa Sanpaolo Innovation Center and led by Luca Cagliero. This project culminated in several publications in both conferences and journals. (2023)
• I was selected to participate in the SAP Business AI Research Retreat, an invitation-only event that brings together leading researchers to discuss SAP's Business AI agenda. My invitation reflects external recognition of the quality and impact of my research, as well as its strategic relevance for collaboration with SAP. (2025)
• This award is comparable to a best paper award at a conference and grants the authors the opportunity to present their work at an A*-ranked conference (NeurIPS, ICLR, or ICML) through the journal-to-conference track. The paper "Think2SQL: Blueprinting Reward Density and Advantage Scaling for Effective Text-to-SQL Reasoning" was accepted for presentation at NeurIPS 2026. (2026)
• Phd Award Template (2026)

Corso di laurea magistrale

• Distributed architectures for big data processing and analytics. A.A. 2025/26, DATA SCIENCE AND ENGINEERING. Collaboratore del corso
• Distributed architectures for big data processing and analytics. A.A. 2023/24, DATA SCIENCE AND ENGINEERING. Collaboratore del corso
• Data science lab: process and methods. A.A. 2023/24, DATA SCIENCE AND ENGINEERING. Collaboratore del corso

Corso di laurea di 1° livello

• Introduction to data science and visualization. A.A. 2025/26, INGEGNERIA INFORMATICA (COMPUTER ENGINEERING). Collaboratore del corso

• Papicchio, S.; Rossi, S.; Cagliero, L.; Papotti, P. (2026)
  Think2SQL: Blueprinting Reward Density and Advantage Scaling for Effective Text-to-SQL Reasoning. In: TRANSACTIONS ON MACHINE LEARNING RESEARCH, vol. 2026-. ISSN 2835-8856
  Contributo su Rivista
• Papicchio, Simone; Cagliero, Luca; Papotti, Paolo (2025)
  SQUAB: Evaluating LLM robustness to Ambiguous and Unanswerable Questions in Semantic Parsing. In: Conference on Empirical Methods in Natural Language Processing, Suzhou (CHN), November 4-9, 2025, pp. 17926-17946. ISBN: 979-8-89176-332-6
  Contributo in Atti di Convegno (Proceeding)
• Papicchio, Simone; Papotti, Paolo; Cagliero, Luca (2025)
  QATCH: Automatic Evaluation of SQL-Centric Tasks on Proprietary Data. In: ACM TRANSACTIONS ON INTELLIGENT SYSTEMS AND TECHNOLOGY, vol. 16, pp. 1-26. ISSN 2157-6904
  Contributo su Rivista
• Feng, Yanlin; Papicchio, Simone; Rahman, Sajjadur (2025)
  CypherBench: Towards Precise Retrieval over Full-scale Modern Knowledge Graphs in the LLM Era. In: Association for Computational Linguistics, Vienna (AT), July 27 - August 1, 2025, pp. 8934-8958. ISBN: 979-8-89176-251-0
  Contributo in Atti di Convegno (Proceeding)
• Papicchio, S.; Papotti, P.; Cagliero, L. (2023)
  QATCH: Benchmarking SQL-centric tasks with Table Representation Learning Models on Your Data. In: 37th Conference on Neural Information Processing Systems, NeurIPS 2023, New Orleans, LA (USA), 2023. ISSN 1049-5258
  Contributo in Atti di Convegno (Proceeding)