News

  • Press SlillZ - RaiPlay My interview where I discuss the impact of AI on the future of work and professions (italian only)
  • Tool TerminatorEconomy.com is a website that enables you to explore the impact of AI on jobs - task by task - by evaluating each task’s level of AI exposure and its potential to be fully automated.
  • Publication Press Towards the Terminator Economy: Assessing Job Exposure to AI through LLMs, with Emilio Colombo, Mario Mezzanzanica and Antonio Serino preprint. Featured in La Repubblica, Radio24, Canadian HRReporter, Wired, and RaiNews24. Just accepted at IJCAI-25!
  • Publication New paper on “ITALIC: An Italian Culture-Aware Natural Language Benchmark” presented at NAACL-25. Featured in La Repubblica
  • Publication New paper on evaluating the proficiency of LLMs on Italian INVALSI benchmark accepted at ECML-PKDD-25.
  • Publication New paper on “RE-FIN: Retrieval-based Enrichment for Financial data” accepted at COLING-25
  • Publication Tool MERLIN, a model agnostic, global, model contrastive explainer for any classifier. Published on Decision Support Systems and available on GitHub
  • Speech “Social Media and AI: No Screen Neededs To Grow Smart”. See my speech about the impact of AI in education (with Pellai, Botturi and Wolf) IT, EN, FR
  • Teaching Students can book a live appointment here

Relevant events I’m involved in as (senior) PC

About

I am Full Professor in Computer Science at University of Milan-Bicocca. I’m director of Master in AI and Data Analytics for Business at University of Milano-Bicocca, Italy.

Current Positions

  • [March24-present] Full Professor at University of Milan-Bicocca
  • [October24-present] Deputy Director Department of Statistics and Quantitative Methods at University of Milan-Bicocca
  • [2022-present] Director of the Master in AI and Data Analytics for Business , Italy
  • [2010-present] Member of the Scientific Committee of CRISP Research Centre, Milan, Italy

Past Positions

  • [Dec2021-Feb2024] Associate Professor at University of Milan-Bicocca
  • [2020-Ago2023] Deputy Director of the CRISP Research Centre, Italy
  • [2016-Nov2021] Assistant Professor at University of Milan-Bicocca
  • [2011-2016] PostDoc at University of Milan-Bicocca
  • [2017-2018] Partner at TabulaeX Ltd (formerly spin-out company of Unimib) working on BI and Big Data Analytics. TabulaeX is now LightCast
  • [2015-2016] Visiting Researcher at King’s College London, AI Planning Group, UK

My research interests include

  • Artificial Intelligence: eXplainable AI, interpretable models, local and global interpreta- tion, explanation through symbolic approaches, fairness.
  • Data Science: Big Data Analytics, ontology Learning, word embedding evaluation, Large Language Models
  • AI Planning [formerly]: domain-independent planning, temporal continuous planning, planning in mixed discrete-continuous domains, planning in hybrid domain

Research team AI@CRISP-UNIMIB

“If you want to go fast go alone. If you want to go far go together”. Proud to work with brilliant phd students and researchers of our team

  • Lorenzo Malandri, Reseacher in AI and Computational Linguistics, Dept of Statistics and Quantitative Methods
  • Navid Nobani, Researcher on XAI, Dept of Statistics and Quantitative Methods at Unimib
  • Francesco Trentini, PostDoc in Economics at Unimib
  • Andrea Seveso, PostDoc in NLP, Dept of Statistics and Quantitative Methods at Unimib
  • Anna Velyka, PostDoc in NLP, Dept of Statistics and Quantitative Methods at Unimib
  • Filippo Pallucchini, PhD Student in Big Data Analytics for Business
  • Simone D’Amico, PhD Student in Big Data Analytics for Business
  • Alessia De Santo, PhD Student in Big Data Analytics for Business
  • Antonio Serino, PhD Student in Big Data Analytics for Business
  • Daniele Potertì, PhD Student in Big Data Analytics for Business
    Past members
  • Anna Giabelli, PhD in Computer Science, now Research Fellow at Istituto Mario Negri
  • Alessandro Castelnovo, PhD in Computer Science, now Data Science Manager at Intesa SanPaolo Bank

(Co)-Developed Tools for Researchers

  • Terminator-Economy is a two-stage AI evaluation pipeline that analyzes the capability of large language models (LLMs), image-processing systems, and robotics to perform specific occupational tasks. The web-app is available at TerminatorEconomy.com. Related paper: IJCAI-25
  • SeNSe is an unsupervised method for embedding alignment via semantic anchor selection. It generates bilingual (or monolingual) word pair dictionaries by selecting “anchors” that share similar contextual neighborhoods across two vocabularies — without assuming prior semantic similarity. Implemented in Python, it enables cross-lingual and cross-domain embedding alignment for NLP research and applications. Related paper: Int. J. Data Sci. Anal. 2025
  • SFAL is a pipeline for Semantic–Functional Alignment Scores. It analyses and links features from sparse autoencoders (SAE) by: (i) calculating co-occurrence matrices of SAE features, (ii) generating embeddings for their explanations (via Neuronpedia), and (iii) scoring the alignment between co-occurrence similarity and semantic similarity of the feature-explanations. Related Paper: EMNLP 2025
  • vec2best is a Python library providing a unified framework for intrinsic evaluation of word-embedding algorithms. It allows researchers to evaluate embedding models using tasks such as similarity, analogy, categorisation and outlier detection, and computes a holistic metric called PCE (Principal Component Evaluation). Related paper: Cognitive Computation 2024
  • MERLIN is a global, model-agnostic, contrastive explainer for any tabular or text classifier. It provides contrastive explanations of how the behaviour of two machine learning models differs. It is available as a python tool on Github. Related papers: Decision Support Systems 2023
  • ContrXT is a model agnostic, global, time contrastive explainer for any text classifier. It is available as a python tool on Github. Related papers: [Information-Fusion-22]
  • TaxoRef is a methodology for Taxonomy Refinement via word embeddings. It allows evaluating the best embedding on the basis of their ability to represent taxonomic similarity relations. Related papers: [ECML-PKDD-21]
  • GraphDBLP is a tool that models (and enriches) DBLP as a graph database for performing graph-based queries and social network analyses. Related papers: [ECML-PKDD-19], [MTAP-18]
  • UPMurphi is a tool for planning with linear and nonlinear continuous PDDL+ models with processes and events. It also handles huge state spaces through a disk-based algorithm. Related papers: [ICAPS-09], [Applied Intelligence 2012]
  • DiNo A Planner built on top of UPMurphi that employs graph-based heuristics to speed-up the plan synthesis (leaded by “Planning Group” at King’s College London). Related papers: [IJCAI-16]

(Co)-Developed LLM Benchmarks

  • ITALIC is a large-scale benchmark dataset of 10K+ multiple-choice questions designed to evaluate natural-language understanding of Italian culture, commonsense reasoning and linguistic proficiency in a morphologically rich language. It spans 12 domains, drawing from real-world public tests and exams. Related paper: NAACL-2025
  • INVALSI-Eval-Suite is a structured benchmark for evaluating large language models’ proficiency on Italian student competencies using the national INVALSI tests. The dataset covers 405 questions across 6 educational grades, spans multiple question-types (multiple choice, multiple complex choice, open-response) and targets both text comprehension and reflection on language. Related paper: ECML-PKDD-25

Granted Ongoing Research Projects

I have been working on the following granted reserch projects, that allow me putting my research on AI and XAI into practice. (see Project)

Education

  • [2012] PhD in Computer Science and Applications Dept of Computer Science, University of L’Aquila, Italy. Advisors: prof. Giuseppe Della Penna (University of L’Aquila, Italy) and prof. Daniele Magazzeni (King’s College London, UK). Topics: AI Planning, Model Checking and Data Quality.
  • [2008] Master Degree in Computer Science and Application University of L’Aquila, Italy. Advisor: prof. Giuseppe Della Penna and prof. Daniele Magazzeni. Topics: AI Planning Control Theory and Model Checking. Maximum score/summa cum laude.

Teaching

  • [2010-present] Lecturer at University of Milano-Bicocca. Topics: Python, SQL, NoSQL Data Stores, Business Intelligence, eXplainable AI, Text Mining
  • [2010-present] Teaching Assistant at University of Milano-Bicocca. Topics: Python, Database, SQL

Please visit my Teaching page for further details

Awards

  • [2019] Research Talent Award I’ve received the first prize at the YoungTalentAward 2019 in collaboration with Accademia Nazionale Lincei in the Computer Science, Engineering & Mathematics area. “for his contribution on applying AI to labour market for describing and predicting labour market phenomena”
  • [2017] FFABR Research Grant for research productivity provided by Italian Ministry of research “Finanziamento annuale individuale delle attività base di ricerca” [Grants provided on a competitive basis aimed at funding research activities]
  • [2014] Best Paper Award at the Third International Conference on Data Technologies and Applications, Vienna, Austria, 29-31, 2014
  • [2013] Best Paper Award at the Third International Workshop, Human Computer Interaction – Knowledge Discovery 1-3 Luglio, Maribor, Slovenia, 2013

Service in International Journals (Selection)

  • (AICom) Associate Editor of AI Communications
    Reviewer for (selection):
  • Artificial Intelligence
  • Cognitive Computation
  • Applied Soft Computing
  • Applied Intelligence
  • Future Generation Computer System
  • Knowledge-Based Systems
  • Computers in Industry
  • Expert Systems with Applications

Program committee membership (Selection)

  • (AAAI) AAAI Conference on Artificial Intelligence (since 2016)
  • (IJCAI) International Joint Conference on Artificial Intelligence (since 2016)
  • (ICAPS) International Conference on Automated Planning and Scheduling (since 2016)
  • (ECML-PKDD) European Conference on Machine Learning and Data Mining (since 2020)
  • ACM/SIGAPP Symposium On Applied Computing (since 2018)
  • Data Technology and Application Conference (since 2014)

Contact

Email (preferred): fabio.mercorio@unimib.it
Office Hours: by appointment

Fabio Mercorio
Room 2043 - U7 Building - Department of Statistics and Quantitative Methods / CRISP Research Centre, University of Milan-Bicocca Viale dell’Innovazione 10 - Milan - Italy Office: (+39) 02 644 82170