Short Résumé

Education

University of Trento

Ph.D. in Computer Science · Trento, IT · Nov. 2023 - Current

• Supervisors: Stefano Teso and Andrea Passerini
• PhD Scholarship: Trustworthy Neuro-Symbolic Machine Learning

University of Trento

Master’s degree in Computer Science · Trento, IT · Sept. 2021 - Oct. 2023

• GPA: 4.0/4.0
• Grade: 110/110 cum laude
• Final dissertation: “From Models to Arguments and Back”, supervised by professors Andrea Passerini and Stefano Teso

University of Trento

Bachelor’s degree in Computer Science · Trento, IT · Sept. 2018 - Jul. 2021

• GPA: 4.0/4.0
• Grade: 110/110 cum laude
• Final dissertation: “Analysis of user warnings in Wikipedia”, supervised by professor Alberto Montresor

Work Experience

Structured Machine Learning Group

Research Intern · Trento, IT · Nov. 2022 - Jun. 2023

• Work on a novel interactive multi-shot debugging protocol that allows the exchange of arguments between a machine and a user in order to correct the model’s beliefs.
• Integrate state-of-the-art eXplainable Artificial Intelligence techniques, such as the ‘Right for the Right Reasons’ loss, into structured prediction output Neuro-Symbolic models like Coherent Hierarchical Multi-label Classification Networks and Semantic Probabilistic Layers.
• Successfully recover the performance of confounded models in the field of hierarchical classification.

Eurecat - Centre Tecnològic de Catalunya

Junior Data Scientist · Barcelona, ES · May. 2021 - Jun. 2021

• Extract Wikipedia data from the Wikipedia dumps - 937GB bz2 and 157GB 7z compressed.
• Design an effective way to extract and analyze the languages spoken by the Wikipedia users - 109’452 database entries.
• Develop a strategy to retrieve Wikipedia users’ User Warnings and Wikibreaks, studying how they affect the users’ activity level - respectively 25’843 and 2’777’181 database entries.
• Build an automated pipeline to download the dumps, extract the data, and compute the statistics using Docker.

Alysso Srl

Junior Software Developer · Trento, IT · Jul. 2017 - Aug. 2017

• Create corporate libraries with the aim of handling database connections regardless of the database management system (SQLite, PostgreSQL, MySQL, and Oracle) in Java.
• Develop a Java-based internal software function that can calculate the distance between two buildings using GIS data.
• Develop a web application in HTML5, CSS3, and JavaScript to show the obtained results.

Social IT

Junior Software Developer · Trento, IT · Jun. 2016 - Jul. 2016

• Contribute to the development of an internal Customer Relationship Management System using Java, JavaScript, HTML5, CSS3, and MySQL.

Teaching

Master in Mechatronics Engineering, University of Trento

Machine Learning FUTURA Tutor · Trento, IT · Sep. 2024 - Dec. 2024

• Introduction to machine learning, probability theory and linear algebra. Standard machine learning algorithms such as decision trees, k-nearest neighbors, Bayesian networks, support vector machines, and kernel machines, along with an introduction to neural networks and deep learning. Model evaluation, parameter estimation, and common techniques in unsupervised and reinforcement learning.

Bachelor in Computer Science, University of Trento

Introduction to Web Programming Teaching Assistant · Trento, IT · Feb. - Jun. 2025; Feb. - Jun. 2026

• Fundamentals of web development, including HTTP protocol, HTML, JavaScript, CSS, request-response cycle, asynchronous JavaScript frameworks (AJAX, AJAJ), state persistence, and database interaction, with practical experience in common web technologies, DOM manipulation, and the design and development of simple web applications using Java Spring.

Master in Artificial Intelligence Systems, University of Trento

Machine Learning Teaching Assistant · Trento, IT · Sep. 2025 - Dec. 2025

• Introduction to machine learning, probability theory and linear algebra. Standard machine learning algorithms such as decision trees, k-nearest neighbors, Bayesian networks, support vector machines, and kernel machines, along with an introduction to neural networks and deep learning. Model evaluation, parameter estimation, and common techniques in unsupervised and reinforcement learning.

Talks

Benchmarking in Neuro-Symbolic AI

Invited Talk · TAILOR, Online · Mar. 2025

Reasoning Shortcuts in Neuro-Symbolic AI

Workshop Talk · IML, University of Edinburgh · Mar. 2026

How to Learn and Not Learn Concepts, Symbols, and Representations

Invited Talk · University of Edinburgh · May. 2026

Research Collaborations

Neuro-Symbolic AI

• Stefano Teso, CiMeC, University of Trento, Italy
• Andrea Passerini, DISI, University of Trento, Italy
• Antonio Vergari, University of Edinburgh, UK
• Emile van Krieken, Vrije Universiteit Amsterdam, The Netherlands
• Han Zhao, University of Illinois Urbana-Champaign, USA
• Efthymia Tsamoura, Huawei Labs, UK
• Tommaso Carraro, Disney Research, Switzerland

Identifiability and Causality

• Emanuele Marconato, CoCaLa, University of Copenhagen

Uncertainty Quantification

• Andrea Pugnana, DISI, University of Trento

Schools & Research Visits

Nordic Probabilistic AI School

Attendee · Copenhagen, DK · Jun. 2024

• Acceptance rate: 18%.
• Topics: Probabilistic models, bayesian workflow, variational inference and optimization, deep generative models, diffusion models, Monte Carlo methods, probabilistic circuits, gaussian processes and causal inference.

School of Informatics, University of Edinburgh

Visiting Researcher · Edinburgh, UK · Mar. 2026 - Jun. 2026

• Visiting the APRIL Lab under the supervision of professor Antonio Vergari.
• Research project on Constraint Learning using Large Language Models.

Software

rsbench: A Neuro-Symbolic Benchmark Suite for Concept Quality and Reasoning Shortcuts

First Author and Maintainer • DeepProbLog, LogicTensorNetworks, CLIP, TCAV, Concept-based Models, Reasoning Shortcuts, #SAT solvers, Blender · NeurIPS 2024, Datasets & Benchmarks · 2024

• Benchmark suite for systematically evaluating reasoning shortcuts in learning & reasoning tasks, covering arithmetic, logic and high-stakes (autonomous driving) settings.
• Configurable data generators (MNMath, MNLogic, Kand-Logic, CLE4EVR, SDD-OIA) for building custom tasks and out-of-distribution splits.
• Concept-quality metrics, plus countrss, a tool that formally verifies whether a task admits reasoning shortcuts by reducing it to model counting, with no training required.
• Code: github.com/unitn-sml/rsbench-code • Data: Zenodo, Website: https://unitn-sml.github.io/rsbench/

auto-nesy-bench: Auto-Formalizing Neuro-Symbolic Predictors

First Author and Maintainer • LLMs, Auto-Formalization, PySAT, CPMpy, PySDD, Sympy, Semantic Probabilistic Layers, Semantic Loss, Knowledge Compilation, Model Counting · Preprint · 2026

• Benchmark for evaluating how well LLMs translate natural-language domain knowledge into logical constraints for Neuro-Symbolic predictors.
• 18 NeSy tasks (9–102 variables) spanning arithmetic, puzzles, ranking, path finding, vision, text and autonomous driving, each with detailed and non-detailed descriptions and annotated data.
• End-to-end pipeline: LLM formalization in five formats (DIMACS, NAT, PySAT, CPMpy, SymPy) with self-verification, compilation to circuits, and training of SPL and Semantic Loss predictors.
• Model-counting-based formula metrics (precision, recall, F1) and downstream metrics (consistency, formula relationship, model-count ratio).
• Code: github.com/unitn-sml/auto-nesy-bench-code • Website: https://unitn-sml.github.io/auto-nesy-bench/

Reviewing

• Uncertainty in Artificial Intelligence (UAI) — 2025, 2026
• Neural Information Processing Systems (NeurIPS) — 2025, 2026
• Unifying Concept Representation Learning Workshop at ICLR — 2026
• Neurosymbolic AI (NeSy) — 2026
• Hybrid Human-Machine Learning and Decision Making at ECML-PKDD — 2026
• Transactions on Machine Learning Research (TMLR) — 2026
• International Conference on Learning Representations (ICLR) — 2026
• Neurosymbolic Artificial Intelligence (NAI) journal — 2026

Skills

• Programming: Python (proficient), Java (proficient), Ruby (intermediate), JavaScript (intermediate), TypeScript (intermediate), R (intermediate), C++ (intermediate), C (intermediate), C# (academic), Matlab (academic)
• Miscellaneous: Linux, Git, GitHub, LaTeX, SQL, PyTorch, TensorFlow, Keras
• Languages: English (C1), Italian (native), German (A2)

Honors & Awards

• Mobility Grant · Erasmus+ Traineeship Programme (2 months) · Barcelona, ES (2021)
• Merit Grant · Premio allo studio Marco Modena (Cassa Rurale Alto Garda - Rovereto, Trento, IT) · Trento, IT (2022)
• Ph.D. scholarship · Three year sponsorship: rank 6th out of 120 participants · Trento, IT (2023)
• Travel Grant · NeurIPS Conference Financial Assistance Award · San Diego, USA (2025)
• Mobility Grant · Erasmus+ Traineeship Programme (3 months) · Edinburgh, UK (2026)

Publications

See the full list on the publications page.