Research

Making AI systems understandable

CRUMBS studies how to explain AI systems, whether they learn from data, reason with symbolic knowledge, or combine the two. Our work is organised around three connected threads.

Learning

Interpretability of learned models: counterfactual and attribution-based explanations for neural networks, robustness of explanations under model changes, and analysis of the internal representations of modern models such as LLMs and diffusion models.

Reasoning

Symbolic and formal methods for explanation: verification of neural networks, computational argumentation, planning, and explanations for sequential decision making.

Explanation

Explanations people can rely on: formal guarantees, contestability and recourse, and applications in high-stakes settings such as finance, energy, and critical infrastructure.

Publications

Curious what this looks like in practice? Here's the full list of what we've been working on, from published papers to our latest pre-prints. Group members are shown in bold.

2026

Journal Probabilistically Robust Counterfactual Explanations under Model Changes L. Marzari, F. Leofante, F. Cicalese, A. Farinelli. Artificial Intelligence, 2026.
Conference Attribution-based Explanations for Markov Decision Processes P. Kobialka, A. Pferscher, F. Leofante, E. Ábrahám, S. Tapia Tarifa, E. Broch Johnsen. Proc. of IJCAI, 2026.
Conference Efficient bias mitigation in T2I diffusion models using Concept Graphs Mansi, A. Kori, F. Leofante. Proc. of NeurIPS, 2026 (to appear).
Conference Enhancing the Robustness of Counterfactual Explanations via Adversarial Training R. Appachi Senthilkumar, F. Leofante, V. Ganesh. Proc. of SAIV, 2026.
Conference Synthesising Counterfactual Explanations via Label-Conditional Gaussian Mixture Variational Autoencoders J. Jiang, F. Leofante, A. Rago, F. Toni. Proc. of ICLR, 2026 (to appear).
Conference What if...Counterfactual Explanations Were to Be Deployed? F. Leofante. Proc. of IJCAI, 2026. PDF

2025

Book Robust Explainable AI Francesco Leofante, Matthew Wicker. Springer, 2025. Paper
Journal Explainable AI, Energy and Critical Infrastructure Systems F. Leofante, A. Artelt, D. Eliades, A. Korre, F. Toni, T. Miller. AI Magazine, 2025.
Journal Explainable AI: definition and attributes of a good explanation for health AI E. Kyrimi et al.. AI Ethics, 2025. PDF
Conference Counterfactual Explanations under Model Multiplicity and their Use in Computational Argumentation G. Alfano, A. Gould, F. Leofante, A. Rago, F. Toni. Proc. of IJCAI, 2025.
Conference Counterfactual Scenarios for Automated Planning N. Gigante, F. Leofante, A. Micheli. Proc. of KR, 2025.
Conference Counterfactual Strategies for Markov Decision Processes P. Kobialka, L. Gerlach, F. Leofante, E. Ábrahám, S. Tapia Tarifa, E. Broch Johnsen. Proc. of IJCAI, 2025.
Conference Representation Consistency for Accurate and Coherent LLM Answer Aggregation J. Jiang, T. Bewley, S. Amoukou, F. Leofante, A. Rago, S. Mishra, F. Toni. Proc. of NeurIPS, 2025.
Conference RobustX: Robust Counterfactual Explanations Made Easy J. Jiang, L. Marzari, A. Purohit, F. Leofante. Proc. of IJCAI, 2025.

2024

Journal Interval Abstractions for Robust Counterfactual Explanations J. Jiang, F. Leofante, A. Rago, F. Toni. Artificial Intelligence, 2024. PDF
Conference Contestable AI needs Computational Argumentation F. Leofante et al.. Proc. of KR, 2024.
Conference Promoting Counterfactual Robustness through Diversity F. Leofante, N. Potyka. Proc. of AAAI, 2024.
Conference Recourse under Model Multiplicity via Argumentative Ensembling J. Jiang, F. Leofante, A. Rago, F. Toni. Proc. of AAMAS, 2024.
Conference Rigorous Probabilistic Guarantees for Robust Counterfactual Explanations L. Marzari, F. Leofante, F. Cicalese, A. Farinelli. Proc. of ECAI, 2024. PDF
Conference Robust Counterfactual Explanations in Machine Learning: A Survey J. Jiang, F. Leofante, A. Rago, F. Toni. Proc. of IJCAI, 2024.
Workshop Explaining Deep Time Series Models in Finance: Are We There Yet? E. Albini, F. Leofante, F. Toni. Proc. of AIFin@ECAI, 2024.

2023

Journal OMTPlan: a Tool for Optimal Planning Modulo Theories F. Leofante. Journal on Satisfiability, Boolean Modeling and Computation, 2023.
Conference Counterfactual Explanations and Model Multiplicity: a Relational Verification View F. Leofante, E. Botoeva, V. Rajani. Proc. of KR, 2023.
Conference Formalising the Robustness of Counterfactual Explanations for Neural Networks J. Jiang, F. Leofante, A. Rago, F. Toni. Proc. of AAAI, 2023.
Conference Provably Robust and Plausible Counterfactual Explanations for Neural Networks via Robust Optimisation J. Jiang, J. Lan, F. Leofante, A. Rago, F. Toni. Proc. of ACML, 2023.
Conference Robust Explanations for Human-Neural Multi-agent Systems with Formal Verification F. Leofante, A. Lomuscio. Proc. of EUMAS, 2023.
Conference Towards Robust Contrastive Explanations for Human-Neural Multi-agent Systems F. Leofante, A. Lomuscio. Proc. of AAMAS, 2023.
Conference Verification of Semantic Key Point Detection for Aircraft Pose Estimation P. Kouvaros, F. Leofante, B. Edwards, C. Chung, D. Margineantu, A. Lomuscio. Proc. of KR, 2023.
Conference Verification-friendly Networks: The Case for Parametric ReLUs F. Leofante, P. Henriksen, A. Lomuscio. Proc. of IJCNN, 2023.

2022

Conference Repairing Misclassifications in Neural Networks Using Limited Data P. Henriksen, F. Leofante, A. Lomuscio. Proc. of SAC, 2022.

2021

Conference Formal Analysis of Neural Network-based Systems in the Aircraft Domain P. Kouvaros, T. Kyono, F. Leofante, A. Lomuscio, D. Margineantu, D. Osipychev, Y. Zheng. Proc. of FM, 2021.
Workshop Robot Swarms as Hybrid Systems: Modelling and Verification S. Schupp, F. Leofante, L. Behr, E. Ábrahám, A. Tacchella. Proc. of SNR@ETAPS, 2021.

2020

Conference Optimal Planning Modulo Theories F. Leofante, E. Giunchiglia, E. Ábrahám, A. Tacchella. Proc. of IJCAI, 2020.
Conference Verification of Neural Networks: Enhancing Scalability through Pruning D. Guidotti, F. Leofante, L. Pulina, A. Tacchella. Proc. of ECAI, 2020.
Workshop ARCH-COMP20 Category Report: Artificial Intelligence and Neural Network Control Systems (AINNCS) for Continuous and Hybrid Systems Plants T. Johnson, D. Manzanas Lopez, P. Musau, H. Tran, E. Botoeva, F. Leofante, A. Maleki, C. Sidrane, J. Fan, C. Huang. Proc. of ARCH Workshop, 2020.
Thesis Optimal Planning Modulo Theories F. Leofante. RWTH Aachen University and University of Genoa, 2020.

2019

Journal Improving Reliability of Myocontrol Using Formal Verification D. Guidotti, F. Leofante, A. Tacchella, C. Castellini. IEEE Transactions on Neural Systems and Rehabilitation Engineering, 2019.
Journal Integrated Synthesis and Execution of Optimal Plans for Multi-Robot Systems in Logistics F. Leofante, E. Ábrahám, T. Niemueller, G. Lakemeyer, A. Tacchella. Information Systems Frontiers, 2019.
Conference Engineering Controllers for Swarm Robotics via Reachability Analysis in Hybrid Systems F. Leofante, S. Schupp, E. Ábrahám, A. Tacchella. Proc. of ECMS, 2019.
Conference Repairing learned controllers with convex optimization: A case study D. Guidotti, F. Leofante, A. Tacchella, C. Castellini. Proc. of CPAIOR, 2019.
Conference SMT-based Planning for Robots in Smart Factories A. Bit-Monnot, F. Leofante, L. Pulina, A. Tacchella. Proc. of IEA/AIE, 2019.
Conference Verification and Repair of Neural Networks: A Progress Report on Convolutional Models D. Guidotti, F. Leofante, L. Pulina, A. Tacchella. Proc. of AI*IA, 2019.
Workshop Reasoning about Neural Networks: a Taxonomy of Current Approaches F. Leofante, N. Narodytska, L. Pulina, A. Tacchella. Proc. of VNN, 2019.

2018

Conference Guaranteed Plans for Multi-Robot Systems via Optimization Modulo Theories F. Leofante. Proc. of AAAI, 2018.
Conference Optimal Multi-robot Task Planning: From Synthesis to Execution (and Back) F. Leofante. Proc. of IJCAI, 2018.
Conference Task Planning with OMT: An Application to Production Logistics F. Leofante, E. Ábrahám, A. Tacchella. Proc. of IFM, 2018.

2017

Conference On the Synthesis of Guaranteed-Quality Plans for Robot Fleets in Logistics Scenarios via Optimization Modulo Theories F. Leofante, E. Ábrahám, T. Niemueller, G. Lakemeyer, A. Tacchella. Proc of IRI, 2017.
Workshop Towards CLIPS-based Task Execution and Monitoring with SMT-based Decision Optimization T. Niemueller, G. Lakemeyer, F. Leofante, E. Ábrahám. Proc. of PlanRob@ICAPS, 2017.

2016

Conference Combining Static and Runtime Methods to Achieve Safe Standing-Up for Humanoid Robots F. Leofante, S. Vuotto, E. Ábrahám, A. Tacchella, N. Jansen. Proc. of ISoLA, 2016.
Conference Learning in Physical Domains: Mating Safety Requirements and Costly Sampling F. Leofante, A. Tacchella. Proc. of AI*IA, 2016.