Alessio Gravina

Modeling graphs demands a careful balance between long-range propagation of information across nodes and the controlled dissipation of noisy or redundant signals to ensure stable learning and generalization. This challenge is exacerbated in dynamic graphs, where structural and temporal information interact, leading to uncontrolled information accumulation and amplifying noise, thereby affecting generalization. We introduce LAMP, a dynamic graph model for snapshot-based dynamic graphs that incorporates adaptive, learned dissipation within a principled dynamical systems framework. Our architecture combines impulsive neural ODEs with an antisymmetric parameterization to model conservative information flow, alongside data-driven dissipative dynamics that regulate information retention over space and time. This formulation yields stable yet expressive representations and enables effective long-range dependency modeling while avoiding pathological information buildup. We provide a theoretical analysis establishing stability guarantees and characterizing the representational power. Extensive experiments on synthetic and real-world benchmarks demonstrate state-of-the-art performance, particularly on tasks requiring extended-range dependency modeling.

@inproceedings{lamp2026,
  title={{Adaptive Memory Retention in Dynamic Graphs}},
  author={Fabrizio De Castelli and Alessio Gravina and Moshe Eliasof and Carola-Bibiane Sch{\"o}nlieb and Davide Bacciu},
  booktitle={Forty-third International Conference on Machine Learning},
  year={2026},
  url={https://openreview.net/forum?id=x7NGYPNlgs}
}

Learning general representations for temporal graphs is challenging because temporal dynamics and structural changes are often intertwined, causing models to capture domain-specific patterns over generalizable ones. This is acting as barrier to the development of foundation models for temporal graphs. Current methods use sequential models alongside Graph Neural Networks (GNNs), producing entangled spatio-temporal representations that struggle to transfer to new graphs. In this work, we hypothesize that this limitation is in part structural. To address this, we propose TIDES, a compositional framework that systematically decouples time from space during learning, serving as a stepping stone toward foundation models for temporal graphs. TIDES relies on a one-time, parallel extraction process to create three domain-agnostic priors: (i) temporal embeddings from a frozen time-series foundation model; (ii) deterministic summaries of local temporal contexts; and (iii) supra-Laplacian positional encodings of the changing graph structure. These features are computed once and used as node inputs for a GNN. TIDES achieves strong performance on multiple benchmarks, reduces negative transfer, and improves cross-graph transferability, outperforming state-of-the-art temporal graph baselines on 20 unseen networks in the MiNT zero-shot benchmark. Overall, TIDES offers a systematic approach to Temporal Graph Foundation Models by providing scalable, transferable representation strategies.

@inproceedings{castelli2026disentangling,
  title={{Disentangling Dynamics: A Compositional Framework for Temporal Graph Foundation Models}},
  author={Fabrizio De Castelli and Alessio Gravina and Davide Bacciu and Moshe Eliasof},
  booktitle={Workshop on Graph Foundation Models: A New Era for Graph Machine Learning},
  year={2026},
  url={https://openreview.net/forum?id=tU97TXAp9x}
}

Effectively capturing long-range interactions remains a fundamental yet unresolved challenge in graph neural network (GNN) research, critical for applications across diverse fields of science. To systematically address this, we introduce ECHO (Evaluating Communication over long HOps), a novel benchmark specifically designed to rigorously assess the capabilities of GNNs in handling very long-range graph propagation. ECHO includes three synthetic graph tasks, namely single-source shortest paths, node eccentricity, and graph diameter, each constructed over diverse and structurally challenging topologies intentionally designed to introduce significant information bottlenecks. ECHO also includes two real-world datasets, ECHO-Charge and ECHO-Energy, which define chemically grounded benchmarks for predicting atomic partial charges and molecular total energies, respectively, with reference computations obtained at the density functional theory (DFT) level. Both tasks inherently depend on capturing complex long-range molecular interactions. Our extensive benchmarking of popular GNN architectures reveals clear performance gaps, emphasizing the difficulty of true long-range propagation and highlighting design choices capable of overcoming inherent limitations. ECHO thereby sets a new standard for evaluating long-range information propagation, also providing a compelling example for its need in AI for science.

@inproceedings{echo2026,
  title={{Can You Hear Me Now? A Benchmark for Long-Range Graph Propagation}},
  author={Luca Miglior and Matteo Tolloso and Alessio Gravina and Davide Bacciu},
  booktitle={The Fourteenth International Conference on Learning Representations},
  year={2026},
  url={https://openreview.net/forum?id=DgkWFPZMPp}
}

Learning to simulate complex physical systems from data has emerged as a promising way to overcome the limitations of traditional numerical solvers, which often require prohibitive computational costs for high-fidelity solutions. Recent Graph Neural Simulators (GNSs) accelerate simulations by learning dynamics on graph-structured data, yet often struggle to capture long-range interactions and suffer from error accumulation under autoregressive rollouts. To address these challenges, we propose Information-preserving Graph Neural Simulators (IGNS), a graph-based neural simulator built on the principles of Hamiltonian dynamics. This structure guarantees preservation of information across the graph, while extending to port-Hamiltonian systems allows the model to capture a broader class of dynamics, including non-conservative effects. IGNS further incorporates a warmup phase to initialize global context, geometric encoding to handle irregular meshes, and a multi-step training objective that facilitates PDE matching, where the trajectory produced by integrating the port-Hamiltonian core aligns with the ground-truth trajectory, thereby reducing rollout error. To evaluate these properties systematically, we introduce new benchmarks that target long-range dependencies and challenging external forcing scenarios. Across all tasks, IGNS consistently outperforms state-of-the-art GNSs, achieving higher accuracy and stability under challenging and complex dynamical systems.

@inproceedings{igns2026,
  title={{Improving Long-Range Interactions in Graph Neural Simulators via Hamiltonian Dynamics}},
  author={Tai Hoang and Alessandro Trenta and Alessio Gravina and Niklas Freymuth and Philipp Becker and Davide Bacciu and Gerhard Neumann},
  booktitle={The Fourteenth International Conference on Learning Representations},
  year={2026},
  url={https://openreview.net/forum?id=x66u6TEDUw}
}

The recent success of State-Space Models (SSMs) in sequence modeling has motivated their adaptation to graph learning, giving rise to Graph State-Space Models (GSSMs). However, existing GSSMs operate by applying SSM modules to sequences extracted from graphs, often compromising core properties such as permutation equivariance, message-passing compatibility, and computational efficiency. In this paper, we introduce a new perspective by embedding the key principles of modern SSM computation directly into the Message-Passing Neural Network framework, resulting in a unified methodology for both static and temporal graphs. Our approach, MP-SSM, enables efficient, permutation-equivariant, and long-range information propagation while preserving the architectural simplicity of message passing. Crucially, MP-SSM enables an exact sensitivity analysis, which we use to theoretically characterize information flow and evaluate issues like vanishing gradients and over-squashing in the deep regime. Furthermore, our design choices allow for a highly optimized parallel implementation akin to modern SSMs. We validate MP-SSM across a wide range of tasks, including node classification, graph property prediction, long-range benchmarks, and spatiotemporal forecasting, demonstrating both its versatility and strong empirical performance.

@article{mpssm2025,
  title={{Message-Passing State-Space Models: Improving Graph Learning with Modern Sequence Modeling}},
  author={Andrea Ceni and Alessio Gravina and Claudio Gallicchio and Davide Bacciu and Carola-Bibiane Schonlieb and Moshe Eliasof},
  year={2025},
  journal={arXiv preprint arXiv:2505.18728}
}

Capturing effective long-range information propagation remains a fundamental yet challenging problem in graph representation learning. Motivated by this, we introduce SONAR, a novel GNN architecture inspired by the dynamics of wave propagation in continuous media. SONAR models information flow on graphs as oscillations governed by the wave equation, allowing it to maintain effective propagation dynamics over long distances. By integrating adaptive edge resistances and state-dependent external forces, our method balances conservative and non-conservative behaviors, improving the ability to learn more complex dynamics. We provide a rigorous theoretical analysis of SONAR's energy conservation and information propagation properties, demonstrating its capacity to address the long-range propagation problem. Extensive experiments on synthetic and real-world benchmarks confirm that SONAR achieves state-of-the-art performance, particularly on tasks requiring long-range information exchange.

@inproceedings{sonar2025,
  title={{SONAR: Long-Range Graph Propagation Through Information Waves}},
  author={Alessandro Trenta and Alessio Gravina and Davide Bacciu},
  booktitle={The Thirty-ninth Annual Conference on Neural Information Processing Systems},
  year={2025},
  url={https://openreview.net/forum?id=Hxfjmc95rl}
}

ChebNet, one of the earliest spectral GNNs, has largely been overshadowed by Message Passing Neural Networks (MPNNs), which gained popularity for their simplicity and effectiveness in capturing local graph structure. Despite their success, MPNNs are limited in their ability to capture long-range dependencies between nodes. This has led researchers to adapt MPNNs through rewiring or make use of Graph Transformers, which compromises the computational efficiency that characterized early spatial message-passing architectures, and typically disregards the graph structure. Almost a decade after its original introduction, we revisit ChebNet to shed light on its ability to model distant node interactions. We find that out-of-box, ChebNet already shows competitive advantages relative to classical MPNNs and GTs on long-range benchmarks, while maintaining good scalability properties for high-order polynomials. However, we uncover that this polynomial expansion leads ChebNet to an unstable regime during training. To address this limitation, we cast ChebNet as a stable and non-dissipative dynamical system, which we coin Stable-ChebNet. Our Stable-ChebNet model allows for stable information propagation, and has controllable dynamics which do not require the use of eigendecompositions, positional encodings, or graph rewiring. Across several benchmarks, Stable-ChebNet achieves near state-of-the-art performance.

@inproceedings{chebnet2025return,
  title={{Return of ChebNet: Understanding and Improving an Overlooked GNN on Long Range Tasks}},
  author={Ali Hariri and Álvaro Arroyo and Alessio Gravina and Moshe Eliasof and Carola-Bibiane Schönlieb and Davide Bacciu and Kamyar Azizzadenesheli and Xiaowen Dong and Pierre Vandergheynst},
  booktitle={The Thirty-ninth Annual Conference on Neural Information Processing Systems},
  year={2025},
  url={https://openreview.net/forum?id=oLyfML1Qze}
}

Graph Neural Networks (GNNs) are models that leverage the graph structure to transmit information between nodes, typically through the message-passing operation. While widely successful, this approach is well-known to suffer from representational collapse as the number of layers increases and insensitivity to the information contained at distant and poorly connected nodes. In this paper, we present a unified view of on the appearance of these issues through the lens of vanishing gradients, using ideas from linear control theory for our analysis. We propose an interpretation of GNNs as recurrent models and empirically demonstrate that a simple state-space formulation of an GNN effectively alleviates these issues at no extra trainable parameter cost. Further, we show theoretically and empirically that (i) Traditional GNNs are by design prone to extreme gradient vanishing even after few layers; (ii) Feature collapse is directly related to the mechanism causing vanishing gradients; (iii) Long-range modeling is most easily achieved by a combination of graph rewiring and vanishing gradient mitigation. We believe our work will help bridge the gap between the recurrent and graph neural network literature and will unlock the design of new deep and performant GNNs.

@inproceedings{gnn-ssm2025,
  title={{On Vanishing Gradients, Over-Smoothing, and Over-Squashing in GNNs: Bridging Recurrent and Graph Learning}},
  author={{\'A}lvaro Arroyo and Alessio Gravina and Benjamin Gutteridge and Federico Barbero and Claudio Gallicchio and Xiaowen Dong and Michael Bronstein and Pierre Vandergheynst},
  booktitle={The Thirty-ninth Annual Conference on Neural Information Processing Systems},
  year={2025},
  url={https://openreview.net/forum?id=N4cyRMuLyl}
}

Graph State Space Models (SSMs) have recently been introduced to enhance Graph Neural Networks (GNNs) in modeling long-range interactions. Despite their success, existing methods either compromise on permutation equivariance or limit their focus to pairwise interactions rather than sequences. Building on the connection between Autoregressive Moving Average (ARMA) and SSM, in this paper, we introduce GRAMA, a Graph Adaptive method based on a learnable ARMA framework that addresses these limitations. By transforming from static to sequential graph data, GRAMA leverages the strengths of the ARMA framework, while preserving permutation equivariance. Moreover, GRAMA incorporates a selective attention mechanism for dynamic learning of ARMA coefficients, enabling efficient and flexible long-range information propagation. We also establish theoretical connections between GRAMA and Selective SSMs, providing insights into its ability to capture long-range dependencies. Experiments on 26 synthetic and real-world datasets demonstrate that GRAMA consistently outperforms backbone models and performs competitively with state-of-the-art methods.

@inproceedings{grama2025,
  title={{Graph Adaptive Autoregressive Moving Average Models}},
  author={Moshe Eliasof and Alessio Gravina and Andrea Ceni and Claudio Gallicchio and Davide Bacciu and Carola-Bibiane Sch{\"o}nlieb},
  booktitle={Forty-second International Conference on Machine Learning},
  year={2025},
  url={https://openreview.net/forum?id=UFlyLkvyAE}
}

Community detection in graphs aims to cluster nodes into meaningful groups, a task particularly challenging in heterophilic graphs, where nodes sharing similarities and membership to the same community are typically distantly connected. This is particularly evident when this task is tackled by graph neural networks, since they rely on an inherently local message passing scheme to learn the node representations that serve to cluster nodes into communities. In this work, we argue that the ability to propagate long-range information during message passing is key to effectively perform community detection in heterophilic graphs. To this end, we introduce the Unsupervised Antisymmetric Graph Neural Network (uAGNN), a novel unsupervised community detection approach leveraging non-dissipative dynamical systems to ensure stability and to propagate long-range information effectively. By employing antisymmetric weight matrices, uAGNN captures both local and global graph structures, overcoming the limitations posed by heterophilic scenarios. Extensive experiments across ten datasets demonstrate uAGNN's superior performance in high and medium heterophilic settings, where traditional methods fail to exploit long-range dependencies. These results highlight uAGNN's potential as a powerful tool for unsupervised community detection in diverse graph environments.

@inproceedings{unsupervised_agnn,
  author={Leeney, William and Gravina, Alessio and Bacciu, Davide},
  booktitle={2025 International Joint Conference on Neural Networks (IJCNN)},
  title={{Non-Dissipative Graph Propagation for Non-Local Community Detection}},
  year={2025},
  pages={1-8},
  doi={10.1109/IJCNN64981.2025.11228363}
}

The dynamics of information diffusion within graphs is a critical open issue that heavily influences graph representation learning, especially when considering long-range propagation. This calls for principled approaches that control and regulate the degree of propagation and dissipation of information throughout the neural flow. Motivated by this, we introduce port-Hamiltonian Deep Graph Networks, a novel framework that models neural information flow in graphs by building on the laws of conservation of Hamiltonian dynamical systems. We reconcile under a single theoretical and practical framework both non-dissipative long-range propagation and non-conservative behaviors, introducing tools from mechanical systems to gauge the equilibrium between the two components. Our approach can be applied to general message-passing architectures, and it provides theoretical guarantees on information conservation in time. Empirical results prove the effectiveness of our port-Hamiltonian scheme in pushing simple graph convolutional architectures to state-of-the-art performance in long-range benchmarks.

@inproceedings{gravina2025phdgn,
  title={Port-Hamiltonian Architectural Bias for Long-Range Propagation in Deep Graph Networks},
  author={Simon Heilig and Alessio Gravina and Alessandro Trenta and Claudio Gallicchio and Davide Bacciu},
  booktitle={The Thirteenth International Conference on Learning Representations},
  year={2025},
  url={https://openreview.net/forum?id=03EkqSCKuO}
}

A common problem in Message-Passing Neural Networks is oversquashing -- the limited ability to facilitate effective information flow between distant nodes. Oversquashing is attributed to the exponential decay in information transmission as node distances increase. This paper introduces a novel perspective to address oversquashing, leveraging dynamical systems properties of global and local non-dissipativity, that enable the maintenance of a constant information flow rate. We present SWAN, a uniquely parameterized GNN model with antisymmetry both in space and weight domains, as a means to obtain non-dissipativity. Our theoretical analysis asserts that by implementing these properties, SWAN offers an enhanced ability to transmit information over extended distances. Empirical evaluations on synthetic and real-world benchmarks that emphasize long-range interactions validate the theoretical understanding of SWAN, and its ability to mitigate oversquashing.

@inproceedings{gravina2025swan,
  title={{On Oversquashing in Graph Neural Networks Through The Lens of Dynamical Systems}},
  author={Alessio Gravina and Moshe Eliasof and Claudio Gallicchio and Davide Bacciu and Carola-Bibiane Sch{\"o}nlieb},
  booktitle={The 39th Annual AAAI Conference on Artificial Intelligence},
  year={2025}
}

Modern graph representation learning works mostly under the assumption of dealing with regularly sampled temporal graph snapshots, which is far from realistic, e.g., social networks and physical systems are characterized by continuous dynamics and sporadic observations. To address this limitation, we introduce the Temporal Graph Ordinary Differential Equation (TG-ODE) framework, which learns both the temporal and spatial dynamics from graph streams where the intervals between observations are not regularly spaced. We empirically validate the proposed approach on several graph benchmarks, showing that TG-ODE can achieve state-of-the-art performance in irregular graph stream tasks.

@inproceedings{gravina2024tgode,
  title     = {{Temporal Graph ODEs for Irregularly-Sampled Time Series}},
  author    = {Gravina, Alessio and Zambon, Daniele and Bacciu, Davide and Alippi, Cesare},
  booktitle = {Proceedings of the Thirty-Third International Joint Conference on Artificial Intelligence, {IJCAI-24}},
  publisher = {International Joint Conferences on Artificial Intelligence Organization},
  editor    = {Kate Larson},
  pages     = {4025--4034},
  year      = {2024},
  month     = {8},
  note      = {Main Track},
  doi       = {10.24963/ijcai.2024/445},
  url       = {https://doi.org/10.24963/ijcai.2024/445}
}

Learning in Continuous-Time Dynamic Graphs (C-TDGs) requires accurately modeling spatio-temporal information on streams of irregularly sampled events. While many methods have been proposed recently, we find that most message passing-, recurrent- or self-attention-based methods perform poorly on long-range tasks. These tasks require correlating information that occurred "far" away from the current event, either spatially (higher-order node information) or along the time dimension (events occurred in the past). To address long-range dependencies, we introduce continuous-time graph anti-symmetric network (CTAN). Grounded within the ordinary differential equations framework, our method is designed for efficient propagation of information. In this paper, we show how CTAN's (i) long-range modeling capabilities are substantiated by theoretical findings and how (ii) its empirical performance on synthetic long-range benchmarks and real-world benchmarks is superior to other methods. Our results motivate CTAN's ability to propagate long-range information in C-TDGs as well as the inclusion of long-range tasks as part of temporal graph models evaluation.

@inproceedings{gravina2024ctan,
  title     = {{Long Range Propagation on Continuous-Time Dynamic Graphs}},
  author    = {Gravina, Alessio and Lovisotto, Giulio and Gallicchio, Claudio and Bacciu, Davide and Grohnfeldt, Claas},
  booktitle = {Proceedings of the 41st International Conference on Machine Learning},
  pages     = {16206--16225},
  year      = {2024},
  volume    = {235},
  series    = {Proceedings of Machine Learning Research},
  publisher = {PMLR},
  url       = {https://proceedings.mlr.press/v235/gravina24a.html}
}

Recent progress in research on deep graph networks (DGNs) has led to a maturation of the domain of learning on graphs. Despite the growth of this research field, there are still important challenges that are yet unsolved. Specifically, there is an urge of making DGNs suitable for predictive tasks on real-world systems of interconnected entities, which evolve over time. With the aim of fostering research in the domain of dynamic graphs, first, we survey recent advantages in learning both temporal and spatial information, providing a comprehensive overview of the current state-of-the-art in the domain of representation learning for dynamic graphs. Second, we conduct a fair performance comparison among the most popular proposed approaches on node-and edge-level tasks, leveraging rigorous model selection and assessment for all the methods, thus establishing a sound baseline for evaluating new architectures and approaches.

@article{gravina2023deep,
  author={Gravina, Alessio and Bacciu, Davide},
  journal={IEEE Transactions on Neural Networks and Learning Systems},
  title={{Deep Learning for Dynamic Graphs: Models and Benchmarks}},
  year={2024},
  pages={1-14},
  doi={10.1109/TNNLS.2024.3379735}
}

Recent research on Deep Graph Networks (DGNs) has broadened the domain of learning on graphs to real-world systems of interconnected entities that evolve over time. This paper addresses prediction problems on graphs defined by a stream of events, possibly irregularly sampled over time, generally referred to as Continuous-Time Dynamic Graphs (C-TDGs). While many predictive problems on graphs may require capturing interactions between nodes at different distances, existing DGNs for C-TDGs are not designed to propagate and preserve long-range information - resulting in suboptimal performance. In this work, we present Continuous-Time Graph Anti-Symmetric Network (CTAN), a DGN for C-TDGs designed within the ordinary differential equations framework that enables efficient propagation of long-range dependencies. We show that our method robustly performs stable and non-dissipative information propagation over dynamically evolving graphs, where the number of ODE discretization steps allows scaling the propagation range. We empirically validate the proposed approach on several real and synthetic graph benchmarks, showing that CTAN leads to improved performance while enabling the propagation of long-range information.

Please refer to the extended version: "Long Range Propagation on Continuous-Time Dynamic Graphs".

Recent advances in Graph Neural Networks (GNNs) have matured the field of learning on graphs, making GNNs essential for prediction tasks in complex, interconnected, and evolving systems. In this paper, we focus on self-supervised, inductive learning for continuous-time dynamic graphs. Without compromising generality, we propose an approach to learn representations and mine anomalies in provenance graphs, which are a form of large-scale, heterogeneous, attributed, and continuous-time dynamic graphs used in the cybersecurity domain, syntactically resembling complex temporal knowledge graphs. We modify the Temporal Graph Network (TGN) framework to heterogeneous input data and directed edges, refining it specifically for inductive learning on provenance graphs. We present and release two pioneering large-scale, continuous-time temporal, heterogeneous, attributed benchmark graph datasets. The datasets incorporate expert-labeled anomalies, promoting subsequent research on representation learning and anomaly detection on intricate real-world networks. Comprehensive experimental analyses of modules, datasets, and baselines underscore the effectiveness of TGN-based inductive learning, affirming its practical utility in identifying semantically significant anomalies in real-world systems.

@inproceedings{reha2023anomaly,
  title={Anomaly Detection in Continuous-Time Temporal Provenance Graphs},
  author={Jakub Reha and Giulio Lovisotto and Michele Russo and Alessio Gravina and Claas Grohnfeldt},
  booktitle={Temporal Graph Learning Workshop @ NeurIPS 2023},
  year={2023},
  url={https://openreview.net/forum?id=88tGIxxhsf}
}

We propose the Hidden Markov Model for temporal Graphs, a deep and fully probabilistic model for learning in the domain of dynamic time-varying graphs. We extend hidden Markov models for sequences to the graph domain by stacking probabilistic layers that perform efficient message passing and learn representations for the individual nodes. We evaluate the goodness of the learned representations on temporal node prediction tasks, and we observe promising results compared to neural approaches.

@inproceedings{hmm_tgl,
  title={Hidden Markov Models for Temporal Graph Representation Learning},
  author={Errica, Federico and Gravina, Alessio and Bacciu, Davide and Micheli, Alessio},
  booktitle={Proceedings of the 31st European Symposium on Artificial Neural Networks, Computational Intelligence and Machine Learning (ESANN)},
  year={2023}
}

Deep Graph Networks (DGNs) currently dominate the research landscape of learning from graphs, due to the efficiency of their adaptive message-passing scheme between nodes. However, DGNs are typically afflicted by a distortion in the information flowing from distant nodes (i.e., over-squashing) that limit their ability to learn long-range dependencies. This reduces their effectiveness, since predictive problems may require to capture interactions at different, and possibly large, radii in order to be effectively solved. We focus on Anti-symmetric Deep Graph Networks (A-DGNs), a recently proposed neural architecture for learning from graphs. A-DGNs are designed based on stable and non-dissipative ordinary differential equations, with a key architectural design based on an anti-symmetric structure of the internal weights. In this paper, we investigate the merits of the resulting architectural bias by incorporating randomized internal connections in node embedding computations and by restricting the training algorithms to operate exclusively at the output layer. To empirically validate our approach, we conduct experiments on various graph benchmarks, demonstrating the effectiveness of the proposed approach in learning from graph data.

@inproceedings{gravina2023randomized,
  author = {Alessio Gravina and Claudio Gallicchio and Davide Bacciu},
  title = {{Non-Dissipative Propagation by Randomized Anti-Symmetric Deep Graph Networks}},
  booktitle = {Machine Learning and Principles and Practice of Knowledge Discovery in Databases},
  year = {2025},
  editor={Meo, Rosa and Silvestri, Fabrizio},
  publisher={Springer Nature Switzerland},
  address={Cham},
  pages={25--36},
  isbn={978-3-031-74643-7}
}

Deep Graph Networks (DGNs) currently dominate the research landscape of learning from graphs, due to their efficiency and ability to implement an adaptive message-passing scheme between the nodes. However, DGNs are typically limited in their ability to propagate and preserve long-term dependencies between nodes, i.e., they suffer from the over-squashing phenomena. This reduces their effectiveness, since predictive problems may require to capture interactions at different, and possibly large, radii in order to be effectively solved. In this work, we present Anti-Symmetric Deep Graph Networks (A-DGNs), a framework for stable and non-dissipative DGN design, conceived through the lens of ordinary differential equations. We give theoretical proof that our method is stable and non-dissipative, leading to two key results: long-range information between nodes is preserved, and no gradient vanishing or explosion occurs in training. We empirically validate the proposed approach on several graph benchmarks, showing that A-DGN leads to improved performance and enables to learn effectively even when dozens of layers are used.

@inproceedings{gravina2023adgn,
  author = {Alessio Gravina and Davide Bacciu and Claudio Gallicchio},
  title = {Anti-Symmetric {DGN}: a stable architecture for Deep Graph Networks},
  booktitle = {The Eleventh International Conference on Learning Representations},
  year = {2023},
  url = {https://openreview.net/forum?id=J3Y7cgZOOS}
}

Deep Graph Networks (DGNs) currently dominate the research landscape of learning from graphs, due to the efficiency of their adaptive message-passing scheme between nodes. However, DGNs are typically limited in their ability to propagate and preserve long-term dependencies between nodes, i.e., they suffer from the over-squashing phenomena. This reduces their effectiveness, since predictive problems may require to capture interactions at different, and possibly large, radii in order to be effectively solved. In this work, we present Anti-Symmetric DGN (A-DGN), a framework for stable and non-dissipative DGN design, conceived through the lens of ordinary differential equations. We give theoretical proof that our method is stable and non-dissipative, leading to two key results: long-range information between nodes is preserved, and no gradient vanishing or explosion occurs in training. We empirically validate the proposed approach on several graph benchmarks, showing that A-DGN yields to improved performance and enables to learn effectively even when dozens of layers are used.

Please refer to the extended version: "Anti-Symmetric DGN: a stable architecture for Deep Graph Networks".

In the early phases of the COVID-19 pandemic, repurposing of drugs approved for use in other diseases helped counteract the aggressiveness of the virus. Therefore, the availability of effective and flexible methodologies to speed up and prioritize the repurposing process is fundamental to tackle present and future challenges to worldwide health. This work addresses the problem of drug repurposing through the lens of deep learning for graphs, by designing an architecture that exploits both structural and biological information to propose a reduced set of drugs that may be effective against an unknown disease. Our main contribution is a method to repurpose a drug against multiple proteins, rather than the most common single-drug/single-protein setting. The method leverages graph embeddings to encode the relevant proteins' and drugs' information based on gene ontology data and structural similarities. Finally, we publicly release a comprehensive and unified data repository for graph-based analysis to foster further studies on COVID-19 and drug repurposing. We empirically validate the proposed approach in a general drug repurposing setting, showing that it generalizes better than single protein repurposing schemes. We conclude the manuscript with an exemplified application of our method to the COVID-19 use case. All source code is publicly available.

@article{gravina2023DrugRep,
  author = {Bacciu, Davide and Errica, Federico and Gravina, Alessio and Madeddu, Lorenzo and Podda, Marco and Stilo, Giovanni},
  title = {Deep Graph Networks for Drug Repurposing with Multi-Protein Targets},
  journal = {IEEE Transactions on Emerging Topics in Computing},
  year = {2023},
  pages={1-14},
  doi={10.1109/TETC.2023.3238963}
}

Schizophrenia is a debilitating psychiatric disorder, leading to both physical and social morbidity. Worldwide 1% of the population is struggling with the disease, with 100,000 new cases annually only in the United States. Despite its importance, the goal of finding effective treatments for schizophrenia remains a challenging task, and previous work conducted expensive large-scale phenotypic screens. This work investigates the benefits of Machine Learning for graphs to optimize drug phenotypic screens and predict compounds that mitigate abnormal brain reduction induced by excessive glial phagocytic activity in schizophrenia subjects. Given a compound and its concentration as input, we propose a method that predicts a score associated with three possible compound effects, i.e., reduce, increase, or not influence phagocytosis. We leverage a high-throughput screening to prove experimentally that our method achieves good generalization capabilities. The screening involves 2218 compounds at five different concentrations. Then, we analyze the usability of our approach in a practical setting, i.e., prioritizing the selection of compounds in the SWEETLEAD library. We provide a list of 64 compounds from the library that have the most potential clinical utility for glial phagocytosis mitigation. Lastly, we propose a novel approach to computationally validate their utility as possible therapies for schizophrenia.

@article{10.1371/journal.pcbi.1009531,
  doi = {10.1371/journal.pcbi.1009531},
  author = {Gravina, Alessio AND Wilson, Jennifer L. AND Bacciu, Davide AND Grimes, Kevin J. AND Priami, Corrado},
  journal = {PLOS Computational Biology},
  publisher = {Public Library of Science},
  title = {Controlling astrocyte-mediated synaptic pruning signals for schizophrenia drug repurposing with deep graph networks},
  year = {2022},
  month = {05},
  volume = {18},
  number = {5},
  pages = {1-19},
  url = {https://doi.org/10.1371/journal.pcbi.1009531}
}

Answer Sentence Selection is one of the steps typically involved in Question Answering. Question Answering is considered a hard task for natural language processing systems, since full solutions would require both natural language understanding and inference abilities. In this paper, we explore how the state of the art in answer selection has improved recently, comparing two of the best proposed models for tackling the problem: the Crossattentive Convolutional Network and the BERT model. The experiments are carried out on two datasets, WikiQA and SelQA, both created for and used in open-domain question answering challenges. We also report on cross domain experiments with the two datasets.

@inproceedings{grs_comparative_study,
  author    = {Gravina, Alessio and Rossetto, Federico and Severini, Silvia and Attardi, Giuseppe},
  editor    = {Bernardi, Raffaella and Navigli, Roberto and Semeraro, Giovanni},
  title     = {A Comparative Study of Models for Answer Sentence Selection},
  booktitle = {Proceedings of the Sixth Italian Conference on Computational Linguistics, Bari, Italy, November 13-15, 2019},
  series    = {{CEUR} Workshop Proceedings},
  volume    = {2481},
  publisher = {CEUR-WS.org},
  year      = {2019},
  url       = {http://ceur-ws.org/Vol-2481/paper64.pdf}
}

Answer Sentence Selection (ASS) is one of the steps typically involved in Question Answering, a hard task for natural language processing since full solutions would require both natural language understanding and world knowledge. We present a new approach to tackle ASS, based on a Cross-Attentive Convolutional Neural Network. The approach was designed for competing in the Fujitsu AI-NLP challenge, which evaluates systems on their performance on the SelQA dataset. This dataset was created on purpose as a benchmark to stress the ability of systems to go beyond simple word co-occurrence criteria. Our submission achieved the top score in the challenge.

@inproceedings{grs_cross_attention,
  author    = {Gravina, Alessio and Rossetto, Federico and Severini, Silvia and Attardi, Giuseppe},
  editor    = {Basile, Pierpaolo and Basile, Valerio and Croce, Danilo and Dell'Orletta, Felice and Guerini, Marco},
  title     = {Cross Attention for Selection-based Question Answering},
  booktitle = {Proceedings of the 2nd Workshop on Natural Language for Artificial Intelligence (NL4AI 2018), Trento, Italy, November 22-23, 2018},
  series    = {{CEUR} Workshop Proceedings},
  volume    = {2244},
  pages     = {53--62},
  publisher = {CEUR-WS.org},
  year      = {2018},
  url       = {http://ceur-ws.org/Vol-2244/paper_05.pdf}
}