Works
Talks
Timeline
CV
Me
(???)
Timeline
2026/01
'Neural Message-Passing on Attention Graphs for Hallucination Detection' has been accepted at ICLR 2026 :) See you in Rio later this year!
2025/11
Really glad to share our LOGML project 'Lost in Serialization: Invariance and Generalization of LLM Graph Reasoners' will be presented at the AAAI 2026 workshop on 'Graphs and more Complex Structures For Learning and Reasoning'!
2025/11
News from AAAI 2026: our 'Beyond Next Token Probabilities' has been accepted as a poster in the main conference track!
2025/10
Extremely glad I have been selected, even this year, amongst NeurIPS' best reviewers!
2025/09
Our paper 'Beyond Token Probes: Hallucination Detection via Activation Tensors with ACT-ViT' has been accepted to NeurIPS 2025! See you in San Diego :)
2025/09
New preprint out: 'Neural Message-Passing on Attention Graphs for Hallucination Detection': we cast hallucination detection in LLMs as a learning task over the heterogeneous, directed graph determined by LLMs' computational traces, including attention maps and activations.
2025/09
Just moderated a super nice GLOW session on 'Oversmoothing, Oversquashing, Heterophily, Long-Range, and more', based on Arnaiz-Rodríguez and Errica's position paper. You can find a recording of the presentation here.
2025/07
LOGML 2025 is over – what an experience it has been! Had the luck to mentor five amazing student on my project, where we explored the invariances, robustness and generalisation abilities of LLMs for graph reasoning. Met delightful people and strong researchers, and had a lot of fun.
2025/07
Had the honour to hold a lecture at Prof. Maron's course on Groups and Deep Learning – I explore the challenges of learning on graphs, focussing on the trade-offs between equivariance, expressive power, and computational complexity, and cover research advances such as random node identifiers, averaging schemes, (sub)structure encodings and the broad paradigm of Subgraph GNNs.
2025/06
Our first GLOW blogpost is out on Substack. Was really a great experience to lead this effort. We summarise thoughts and opinions on the past, present and future of Graph Learning reaserch as emerging from our dedicated GLOW sessions.
2025/05
Good news: two of our works have been accepted at ICML 2025: 'Balancing Efficiency and Expressiveness: Subgraph GNNs with Walk-Based Centrality' and 'Graph Learning Will Lose Relevance Due To Poor Benchmarks'. I will be in Vancouver presenting them :)
2025/05
Our paper Understanding and Improving Laplacian Positional Encodings For Temporal GNNs got accepted at ECML-PKDD 2025! We contribute an exploration and improved understanding of Laplacian Positional Encodings for Temporal GNNs, along wit a faster, scalable implementation :)
2025/03
I have been selected as a mentor at this year's LOGML Summer School! Super excited to take part in July :)
2025/03
Our new work Learning on LLM Output Signatures for Gray-Box Behavior Analysis is being featured as a spotlight contribution at the ICLR 2025 Workshop on 'Quantifying Uncertainty and Hallucination in Foundation Models'! We study the problems of automated data contamination and hallucination detection in LLMs. We propose a unified gray-box approach which learns on LLM output signatures (LOS), achieving state-of-the-art performance and strong cross-model transfer.
2025/02
Just released a new preprint: a position paper where we discuss the problematic aspects of current benchmarking datasets and practices in Graph Learning. We experimentally validate our claims and propose possible solutions.
2025/02
We have just held an amazing GLOW session today. More than fifty researchers gathered to discuss together the past, present and future of Machine Learning on graphs. It was super fun to moderate it :D
2025/01
New preprint out: Balancing Efficiency and Expressiveness: Subgraph GNNs with Walk-Based Centrality – we employ walk-based centrality measures for subgraph sampling to drastically speed-up Subgraph GNNs and retain discriminative power!
2025/01
Our paper Topological Blindspots: Understanding and Extending Topological Deep Learning Through the Lens of Expressivity has been selected for an oral presentation at ICLR 2025 :)
2024/12
I have just held a Keynote Talk at the Italian LoG 2024 Meetup in Siena on my works related to Subgraph GNNs. What a nice atmosphere and great community.
2024/12
Our Flexible, Equivariant Framework for Subgraph GNNs via Graph Products and Graph Coarsening received the Best Paper Award at the NeurIPS 2024 NeurReps workshop, after being presented as an oral!
2024/11
I am honoured to have been selected as Top Reviewer at NeurIPS 2024!
2024/10
I am co-organising GLOW (Graph Learning On Wednesdays), a new monthly Graph Learning reading group starting in October 2024. Happy to explore new formats and topics to bring our community together and discuss where we are heading!
2024/10
Our paper Towards Foundation Models on Graphs: An Analysis on Cross-Dataset Transfer of Pretrained GNNs will be presented at the NeurIPS 2024 workshop on Symmetry and Geometry in Neural Representations (NeurReps). We experimentally explore the extent to which pretrained Graph Neural Networks can be applied across datasets, an effort requiring to be agnostic to dataset-specific features and their encodings.
2024/09
Happy to share our paper Flexible, Equivariant Framework for Subgraph GNNs via Graph Products and Graph Coarsening has been accepted as a poster at NeurIPS 2024 :)
2024/08
Excited to be invited as a Keynote Speaker at the Italian LoG 2024 Meetup in Siena happening in December :)
2024/07
I just took part as a panelist at the ICML 2024 Graph Machine Learning social -- very interesting discussions on the new emerging directions of Graph Learning :)
2024/05
Our position paper Future Directions in the Theory of Graph Machine Learning has been accepted at ICML 2024 :) Come at the poster session to discuss the current status of theoretical advancements for learning on graphs.
2024/03
I am excited to share I will be spending some time in Vienna! I will be visiting Prof. Thomas Gärtners' Lab at TU Wien. :)
2024/01
Quite an update: I have just started as a Postdoctoral Fellow researcher at Technion (Israel), where I will work with Prof. Haggai Maron on Equivariance, Expressiveness, Graph Neural Networks, Geometric Deep Learning and friends :)
2023/11
Happy to share 'Edge Directionality Improves Learning on Heterophilic Graphs' has been accepted to presented at the LoG 2023 conference!
2023/11
Big news: I have just discussed my PhD thesis with examiners Prof. Ben Glocker and Prof. Yaron Lipman. I am excited to share I have passed the viva examination without corrections, this marking the (successful) end of my PhD journey :D
2023/08
It's been a while, huh? ... Well, these past months I've been mostly working on my PhD thesis 'Expressive and Efficient Graph Neural Networks'. I am happy to share that I have finally submitted it!
2023/05
New preprint out: Edge Directionality Improves Learning on Heterophilic Graphs. We study how to handle directed graphs in a principled manner (and what are the theoretical and practical benefits of the approach, especially in heterophilic settings).
2023/04
Our paper Graph Positional Encoding via Random Feature Propagation has been accepted to ICML 2023!
2023/03
I had the honour to give a talk as the penultimate lecture of the Geometric Deep Learning course at Oxford University. A great experience with a lot of interesting questions from the students :)
2023/01
Our paper Graph Neural Networks for Link Prediction with Subgraph Sketching has been accepted at ICLR 2023 as a notable, top 5% paper!
2023/01
The video of our tutorial 'Exploring the practical and theoretical landscape of expressive Graph Neural Networks' is available on youtube :)
2022/12
Just held our tutorial at the Learning on Graph conference. Amazing experience with a lot of emerging interesting discussions!
2022/12
I have been selected as one of the twenty best reviewers at the first edition of the Learning on Graph conference! I have also gathered some reflections and unsolicited thoughts in a short presentation I held during the opening remarks, take a look :)
2022/10
Our tutorial proposal got accepted at Learning on Graph 2022! The tutorial will focus on the expressive power of Graph Neural Networks and I will present along with Beatrice Bevilacqua and Dr. Haggai Maron.
2022/10
Understanding and Extending Subgraph GNNs by Rethinking Their Symmetries got selected as oral presentation at NeurIPS 2022!
2022/09
Our new paper Understanding and Extending Subgraph GNNs by Rethinking Their Symmetries is accepted at NeurIPS 2022 – see you in New Orleans :)
2022/07
I am honoured I will contribute to the Geometric Deep Learning AIMS 2022 course by holding a seminar on 'Subgraphs for more expressive GNNs'.
2022/07
Exciting news – I got selected to join the LOGML 2022 summer school and I will work with Leonardo Cotta and Dr. Shubhendu Trivedi on 'Equivariant Poset Representations'.
2022/06
New blogpost out! It covers the two works Cris Bodnar and I authored on 'Topological Message Passing'.
2022/06
A new blogpost from Airbnb explains how they chose our SIGN model to suit their graph learning needs at scale.
2022/03
Just given a talk at the Dagstuhl Seminar 'Graph Embeddings: Theory meets Practice'. I presented '(Graph) Representation Learning on Simplicial and Cellular Complexes' and spent three days with amazing researchers.
2022/03
A new interview by Zak Jost just out. It covers our recent paper 'Equivariant Subgraph Aggregation Networks.
2022/01
Our paper 'Equivariant Subgraph Aggregation Networks' has been accepted at ICLR 2022 as a spotlight!
2021/12
I've just presented 'Subgraphs for more expressive GNNs' at the 2021 Nepal Winter School. Honoured!
2021/12
New co-authored blogpost out: 'Using Subgraphs for More Expressive GNNs'. Gist: we gather together with M. M. Bronstein, C. Morris, L. Cotta, H. Maron and L. Zhao and discuss our related concurrent works together.
2021/11
I am glad to share our paper 'Weisfeiler and Lehman Go Cellular: CW Networks' got accepted at NeurIPS 2021!
2021/10
I had the pleasure to moderate the London ML Meetup on the latest works from Dr. Anees Kazi on GNNs for medical applications.
2021/09
Excited to share I will spend October and November in Spain (València & Málaga). I will continue my research work remotely during this period :)
2021/08
The code for simplicial and cellular message-passing which powers our latest approaches is out now!
2021/08
I got selected to take part in the LOGML 2021 summer school! I will join Dr. Haggai Maron on exploring subgraphs for more expressive GNNs.
2021/07
Our latest paper on cellular message-passing was featured in a blogpost from Synced!
2021/06
Just gave a talk with Cris Bodnar at TopoNets 2021 on our latest works on topological message-passing.
2021/06
New pre-print out: 'Weisfeiler and Lehman Go Cellular: CW Networks'. We continue our exploration of message-passing on more general topological objects.
2021/06
New post out on 'Provably expressive Graph Neural Network' on the Twitter Engineering blog :)
2021/05
Cris Bodnar and I are presenting our paper 'Weisfeiler and Lehman Go Topological: Message Passing Simplicial Networks' at Cambridge's AI Research Talks :)
2021/05
Our paper Weisfeiler and Lehman Go Topological: Message Passing Simplicial Networks' got accepted at ICML 2021!
2021/04
I have presented 'Weisfeiler and Lehman Go Topological: Message Passing Simplicial Networks' along with Cris Bodnar and Yu Guang Wang at the MPI MIS + UCLA Math Machine Learning Seminar :)
2021/04
Weisfeiler and Lehman Go Topological: Message Passing Simplicial Networks' got accepted as a spotlight at the GTRL ICLR 2021 workshop!
2021/03
New pre-print out: 'Weisfeiler and Lehman Go Topological: Message Passing Simplicial Networks'. We explore generalisations of message-passing from graphs to simplicial complexes.
2020/10
Happy to share I am starting (again) as a PhD candidate at Imperial College London under the supervision of Prof. Michael Bronstein. I will conduct my research activity part-time while working for Twitter.
2020/04
New work on arXiv: 'SIGN: Scalable Inception Graph Neural Networks'. We propose a simple, yet effective approach to scale the computation of Graph Neural Networks to large-scale graphs.
2019/12
I will be at NeurIPS 2019 in the following days to present our paper 'Learning Interpretable Disease Self-Representations for Drug Repositioning' at the NeurIPS 2019 Graph Representation Learning Workshop :)
2019/09
New pre-print out: 'Learning Interpretable Disease Self-Representations for Drug Repositioning'. We train linear sparse self-encoding models for repurposing drugs.
2019/06
Thrilled to announce Fabula AI is being acquired by Twitter and I will join on-board as a tweep Machine Learning Engineer!
2019/05
I will spend one week in LA to attend the IPAM workshop 'Deep Geometric Learning of Big Data and Applications' – see you there :)
2018/10
I am starting a PhD at USI (Università della Svizzera Italiana) advised by Prof. Bronstein. However, I will spend my first 4 months in London visting him at Imperial College London!
2018/06
After graduating with distinction from Politecnico di Milano, I am happy to share that I have joined startup Fabula AI to explore applications of Geometric Deep Learning to Fake News Detection :)
My personal website has been realised through Jekyll and Github Pages. The theme — slightly customised — is by orderedlist. Drop me a message if you'd fancy a chat!