Davide Bacciu

Andrea Cossu, Tinne Tuytelaars, Antonio Carta et al.

Pre-trained models are nowadays a fundamental component of machine learning research. In continual learning, they are commonly used to initialize the model before training on the stream of non-stationary data. However, pre-training is rarely applied during continual learning. We formalize and investigate the characteristics of the continual pre-training scenario in both language and vision environments, where a model is continually pre-trained on a stream of incoming data and only later fine-tuned to different downstream tasks. We show that continually pre-trained models are robust against catastrophic forgetting and we provide strong empirical evidence supporting the fact that self-supervised pre-training is more effective in retaining previous knowledge than supervised protocols. Code is provided at https://github.com/AndreaCossu/continual-pretraining-nlp-vision .

Hamed Hemati, Andrea Cossu, Antonio Carta et al. · berkeley

Real-world data streams naturally include the repetition of previous concepts. From a Continual Learning (CL) perspective, repetition is a property of the environment and, unlike replay, cannot be controlled by the agent. Nowadays, the Class-Incremental (CI) scenario represents the leading test-bed for assessing and comparing CL strategies. This scenario type is very easy to use, but it never allows revisiting previously seen classes, thus completely neglecting the role of repetition. We focus on the family of Class-Incremental with Repetition (CIR) scenario, where repetition is embedded in the definition of the stream. We propose two stochastic stream generators that produce a wide range of CIR streams starting from a single dataset and a few interpretable control parameters. We conduct the first comprehensive evaluation of repetition in CL by studying the behavior of existing CL strategies under different CIR streams. We then present a novel replay strategy that exploits repetition and counteracts the natural imbalance present in the stream. On both CIFAR100 and TinyImageNet, our strategy outperforms other replay approaches, which are not designed for environments with repetition.

Hamed Hemati, Vincenzo Lomonaco, Davide Bacciu et al. · berkeley

Hypernetworks mitigate forgetting in continual learning (CL) by generating task-dependent weights and penalizing weight changes at a meta-model level. Unfortunately, generating all weights is not only computationally expensive for larger architectures, but also, it is not well understood whether generating all model weights is necessary. Inspired by latent replay methods in CL, we propose partial weight generation for the final layers of a model using hypernetworks while freezing the initial layers. With this objective, we first answer the question of how many layers can be frozen without compromising the final performance. Through several experiments, we empirically show that the number of layers that can be frozen is proportional to the distributional similarity in the CL stream. Then, to demonstrate the effectiveness of hypernetworks, we show that noisy streams can significantly impact the performance of latent replay methods, leading to increased forgetting when features from noisy experiences are replayed with old samples. In contrast, partial hypernetworks are more robust to noise by maintaining accuracy on previous experiences. Finally, we conduct experiments on the split CIFAR-100 and TinyImagenet benchmarks and compare different versions of partial hypernetworks to latent replay methods. We conclude that partial weight generation using hypernetworks is a promising solution to the problem of forgetting in neural networks. It can provide an effective balance between computation and final test accuracy in CL streams.

Alessio Gravina, Davide Bacciu, Claudio Gallicchio

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 yields to improved performance and enables to learn effectively even when dozens of layers are used.

Antonio Carta, Andrea Cossu, Vincenzo Lomonaco et al.

Distributed learning on the edge often comprises self-centered devices (SCD) which learn local tasks independently and are unwilling to contribute to the performance of other SDCs. How do we achieve forward transfer at zero cost for the single SCDs? We formalize this problem as a Distributed Continual Learning scenario, where SCD adapt to local tasks and a CL model consolidates the knowledge from the resulting stream of models without looking at the SCD's private data. Unfortunately, current CL methods are not directly applicable to this scenario. We propose Data-Agnostic Consolidation (DAC), a novel double knowledge distillation method that consolidates the stream of SC models without using the original data. DAC performs distillation in the latent space via a novel Projected Latent Distillation loss. Experimental results show that DAC enables forward transfer between SCDs and reaches state-of-the-art accuracy on Split CIFAR100, CORe50 and Split TinyImageNet, both in reharsal-free and distributed CL scenarios. Somewhat surprisingly, even a single out-of-distribution image is sufficient as the only source of data during consolidation.

Mattia Sangermano, Antonio Carta, Andrea Cossu et al.

Online Continual learning is a challenging learning scenario where the model must learn from a non-stationary stream of data where each sample is seen only once. The main challenge is to incrementally learn while avoiding catastrophic forgetting, namely the problem of forgetting previously acquired knowledge while learning from new data. A popular solution in these scenario is to use a small memory to retain old data and rehearse them over time. Unfortunately, due to the limited memory size, the quality of the memory will deteriorate over time. In this paper we propose OLCGM, a novel replay-based continual learning strategy that uses knowledge condensation techniques to continuously compress the memory and achieve a better use of its limited size. The sample condensation step compresses old samples, instead of removing them like other replay strategies. As a result, the experiments show that, whenever the memory budget is limited compared to the complexity of the data, OLCGM improves the final accuracy compared to state-of-the-art replay strategies.

Alessio Gravina, Davide Bacciu

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 realworld systems of interconnected entities, which evolve over time. With the aim of fostering research in the domain of dynamic graphs, at 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. Secondly, 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

Gabriele Merlin, Vincenzo Lomonaco, Andrea Cossu et al.

Continual Learning requires the model to learn from a stream of dynamic, non-stationary data without forgetting previous knowledge. Several approaches have been developed in the literature to tackle the Continual Learning challenge. Among them, Replay approaches have empirically proved to be the most effective ones. Replay operates by saving some samples in memory which are then used to rehearse knowledge during training in subsequent tasks. However, an extensive comparison and deeper understanding of different replay implementation subtleties is still missing in the literature. The aim of this work is to compare and analyze existing replay-based strategies and provide practical recommendations on developing efficient, effective and generally applicable replay-based strategies. In particular, we investigate the role of the memory size value, different weighting policies and discuss about the impact of data augmentation, which allows reaching better performance with lower memory sizes.

Danilo Numeroso, Davide Bacciu, Petar Veličković

Neural Algorithmic Reasoning is an emerging area of machine learning which seeks to infuse algorithmic computation in neural networks, typically by training neural models to approximate steps of classical algorithms. In this context, much of the current work has focused on learning reachability and shortest path graph algorithms, showing that joint learning on similar algorithms is beneficial for generalisation. However, when targeting more complex problems, such similar algorithms become more difficult to find. Here, we propose to learn algorithms by exploiting duality of the underlying algorithmic problem. Many algorithms solve optimisation problems. We demonstrate that simultaneously learning the dual definition of these optimisation problems in algorithmic learning allows for better learning and qualitatively better solutions. Specifically, we exploit the max-flow min-cut theorem to simultaneously learn these two algorithms over synthetically generated graphs, demonstrating the effectiveness of the proposed approach. We then validate the real-world utility of our dual algorithmic reasoner by deploying it on a challenging brain vessel classification task, which likely depends on the vessels' flow properties. We demonstrate a clear performance gain when using our model within such a context, and empirically show that learning the max-flow and min-cut algorithms together is critical for achieving such a result.

Riccardo Massidda, Atticus Geiger, Thomas Icard et al.

Causal abstraction provides a theory describing how several causal models can represent the same system at different levels of detail. Existing theoretical proposals limit the analysis of abstract models to "hard" interventions fixing causal variables to be constant values. In this work, we extend causal abstraction to "soft" interventions, which assign possibly non-constant functions to variables without adding new causal connections. Specifically, (i) we generalize $τ$-abstraction from Beckers and Halpern (2019) to soft interventions, (ii) we propose a further definition of soft abstraction to ensure a unique map $ω$ between soft interventions, and (iii) we prove that our constructive definition of soft abstraction guarantees the intervention map $ω$ has a specific and necessary explicit form.

Daniele Atzeni, Federico Errica, Davide Bacciu et al.

We propose an extension of the Contextual Graph Markov Model, a deep and probabilistic machine learning model for graphs, to model the distribution of edge features. Our approach is architectural, as we introduce an additional Bayesian network mapping edge features into discrete states to be used by the original model. In doing so, we are also able to build richer graph representations even in the absence of edge features, which is confirmed by the performance improvements on standard graph classification benchmarks. Moreover, we successfully test our proposal in a graph regression scenario where edge features are of fundamental importance, and we show that the learned edge representation provides substantial performance improvements against the original model on three link prediction tasks. By keeping the computational complexity linear in the number of edges, the proposed model is amenable to large-scale graph processing.

Gabriele Lagani, Davide Bacciu, Claudio Gallicchio et al.

Features extracted from Deep Neural Networks (DNNs) have proven to be very effective in the context of Content Based Image Retrieval (CBIR). In recent work, biologically inspired \textit{Hebbian} learning algorithms have shown promises for DNN training. In this contribution, we study the performance of such algorithms in the development of feature extractors for CBIR tasks. Specifically, we consider a semi-supervised learning strategy in two steps: first, an unsupervised pre-training stage is performed using Hebbian learning on the image dataset; second, the network is fine-tuned using supervised Stochastic Gradient Descent (SGD) training. For the unsupervised pre-training stage, we explore the nonlinear Hebbian Principal Component Analysis (HPCA) learning rule. For the supervised fine-tuning stage, we assume sample efficiency scenarios, in which the amount of labeled samples is just a small fraction of the whole dataset. Our experimental analysis, conducted on the CIFAR10 and CIFAR100 datasets shows that, when few labeled samples are available, our Hebbian approach provides relevant improvements compared to various alternative methods.

Julio Hurtado, Alain Raymond-Saez, Vladimir Araujo et al.

Catastrophic forgetting, the phenomenon of forgetting previously learned tasks when learning a new one, is a major hurdle in developing continual learning algorithms. A popular method to alleviate forgetting is to use a memory buffer, which stores a subset of previously learned task examples for use during training on new tasks. The de facto method of filling memory is by randomly selecting previous examples. However, this process could introduce outliers or noisy samples that could hurt the generalization of the model. This paper introduces Memory Outlier Elimination (MOE), a method for identifying and eliminating outliers in the memory buffer by choosing samples from label-homogeneous subpopulations. We show that a space with a high homogeneity is related to a feature space that is more representative of the class distribution. In practice, MOE removes a sample if it is surrounded by samples from different labels. We demonstrate the effectiveness of MOE on CIFAR-10, CIFAR-100, and CORe50, outperforming previous well-known memory population methods.

Davide Bacciu, Alessio Conte, Francesco Landolfi

Downsampling produces coarsened, multi-resolution representations of data and it is used, for example, to produce lossy compression and visualization of large images, reduce computational costs, and boost deep neural representation learning. Unfortunately, due to their lack of a regular structure, there is still no consensus on how downsampling should apply to graphs and linked data. Indeed reductions in graph data are still needed for the goals described above, but reduction mechanisms do not have the same focus on preserving topological structures and properties, while allowing for resolution-tuning, as is the case in regular data downsampling. In this paper, we take a step in this direction, introducing a unifying interpretation of downsampling in regular and graph data. In particular, we define a graph coarsening mechanism which is a graph-structured counterpart of controllable equispaced coarsening mechanisms in regular data. We prove theoretical guarantees for distortion bounds on path lengths, as well as the ability to preserve key topological properties in the coarsened graphs. We leverage these concepts to define a graph pooling mechanism that we empirically assess in graph classification tasks, providing a greedy algorithm that allows efficient parallel implementation on GPUs, and showing that it compares favorably against pooling methods in literature.

Rudy Semola, Vincenzo Lomonaco, Davide Bacciu

Predictive machine learning models nowadays are often updated in a stateless and expensive way. The two main future trends for companies that want to build machine learning-based applications and systems are real-time inference and continual updating. Unfortunately, both trends require a mature infrastructure that is hard and costly to realize on-premise. This paper defines a novel software service and model delivery infrastructure termed Continual Learning-as-a-Service (CLaaS) to address these issues. Specifically, it embraces continual machine learning and continuous integration techniques. It provides support for model updating and validation tools for data scientists without an on-premise solution and in an efficient, stateful and easy-to-use manner. Finally, this CL model service is easy to encapsulate in any machine learning infrastructure or cloud system. This paper presents the design and implementation of a CLaaS instantiation, called LiquidBrain, evaluated in two real-world scenarios. The former is a robotic object recognition setting using the CORe50 dataset while the latter is a named category and attribute prediction using the DeepFashion-C dataset in the fashion domain. Our preliminary results suggest the usability and efficiency of the Continual Learning model services and the effectiveness of the solution in addressing real-world use-cases regardless of where the computation happens in the continuum Edge-Cloud.

Emanuele Cosenza, Andrea Valenti, Davide Bacciu

Graphs can be leveraged to model polyphonic multitrack symbolic music, where notes, chords and entire sections may be linked at different levels of the musical hierarchy by tonal and rhythmic relationships. Nonetheless, there is a lack of works that consider graph representations in the context of deep learning systems for music generation. This paper bridges this gap by introducing a novel graph representation for music and a deep Variational Autoencoder that generates the structure and the content of musical graphs separately, one after the other, with a hierarchical architecture that matches the structural priors of music. By separating the structure and content of musical graphs, it is possible to condition generation by specifying which instruments are played at certain times. This opens the door to a new form of human-computer interaction in the context of music co-creation. After training the model on existing MIDI datasets, the experiments show that the model is able to generate appealing short and long musical sequences and to realistically interpolate between them, producing music that is tonally and rhythmically consistent. Finally, the visualization of the embeddings shows that the model is able to organize its latent space in accordance with known musical concepts.

Lorenzo Simone, Davide Bacciu

High-quality synthetic data can support the development of effective predictive models for biomedical tasks, especially in rare diseases or when subject to compelling privacy constraints. These limitations, for instance, negatively impact open access to electrocardiography datasets about arrhythmias. This work introduces a self-supervised approach to the generation of synthetic electrocardiography time series which is shown to promote morphological plausibility. Our model (ECGAN) allows conditioning the generative process for specific rhythm abnormalities, enhancing synchronization and diversity across samples with respect to literature models. A dedicated sample quality assessment framework is also defined, leveraging arrhythmia classifiers. The empirical results highlight a substantial improvement against state-of-the-art generative models for sequences and audio synthesis.

Federico Matteoni, Andrea Cossu, Claudio Gallicchio et al.

Continual Learning (CL) on time series data represents a promising but under-studied avenue for real-world applications. We propose two new CL benchmarks for Human State Monitoring. We carefully designed the benchmarks to mirror real-world environments in which new subjects are continuously added. We conducted an empirical evaluation to assess the ability of popular CL strategies to mitigate forgetting in our benchmarks. Our results show that, possibly due to the domain-incremental properties of our benchmarks, forgetting can be easily tackled even with a simple finetuning and that existing strategies struggle in accumulating knowledge over a fixed, held-out, test subject.

Valerio De Caro, Claudio Gallicchio, Davide Bacciu

We propose a novel algorithm for performing federated learning with Echo State Networks (ESNs) in a client-server scenario. In particular, our proposal focuses on the adaptation of reservoirs by combining Intrinsic Plasticity with Federated Averaging. The former is a gradient-based method for adapting the reservoir's non-linearity in a local and unsupervised manner, while the latter provides the framework for learning in the federated scenario. We evaluate our approach on real-world datasets from human monitoring, in comparison with the previous approach for federated ESNs existing in literature. Results show that adapting the reservoir with our algorithm provides a significant improvement on the performance of the global model.

Francesco Corti, Rahim Entezari, Sara Hooker et al.

We study the impact of different pruning techniques on the representation learned by deep neural networks trained with contrastive loss functions. Our work finds that at high sparsity levels, contrastive learning results in a higher number of misclassified examples relative to models trained with traditional cross-entropy loss. To understand this pronounced difference, we use metrics such as the number of PIEs (Hooker et al., 2019), Q-Score (Kalibhat et al., 2022), and PD-Score (Baldock et al., 2021) to measure the impact of pruning on the learned representation quality. Our analysis suggests the schedule of the pruning method implementation matters. We find that the negative impact of sparsity on the quality of the learned representation is the highest when pruning is introduced early on in the training phase.

Andrea Valenti, Davide Bacciu

Disentanglement is a difficult property to enforce in neural representations. This might be due, in part, to a formalization of the disentanglement problem that focuses too heavily on separating relevant factors of variation of the data in single isolated dimensions of the neural representation. We argue that such a definition might be too restrictive and not necessarily beneficial in terms of downstream tasks. In this work, we present an alternative view over learning (weakly) disentangled representations, which leverages concepts from relational learning. We identify the regions of the latent space that correspond to specific instances of generative factors, and we learn the relationships among these regions in order to perform controlled changes to the latent codes. We also introduce a compound generative model that implements such a weak disentanglement approach. Our experiments shows that the learned representations can separate the relevant factors of variation in the data, while preserving the information needed for effectively generating high quality data samples.

Andrea Cossu, Francesco Spinnato, Riccardo Guidotti et al.

Continual Learning trains models on a stream of data, with the aim of learning new information without forgetting previous knowledge. Given the dynamic nature of such environments, explaining the predictions of these models can be challenging. We study the behavior of SHAP values explanations in Continual Learning and propose an evaluation protocol to robustly assess the change of explanations in Class-Incremental scenarios. We observed that, while Replay strategies enforce the stability of SHAP values in feedforward/convolutional models, they are not able to do the same with fully-trained recurrent models. We show that alternative recurrent approaches, like randomized recurrent models, are more effective in keeping the explanations stable over time.

Andrea Valenti, Davide Bacciu, Antonio Vergari

While showing impressive performance on various kinds of learning tasks, it is yet unclear whether deep learning models have the ability to robustly tackle reasoning tasks. than by learning the underlying reasoning process that is actually required to solve the tasks. Measuring the robustness of reasoning in machine learning models is challenging as one needs to provide a task that cannot be easily shortcut by exploiting spurious statistical correlations in the data, while operating on complex objects and constraints. reasoning task. To address this issue, we propose ChemAlgebra, a benchmark for measuring the reasoning capabilities of deep learning models through the prediction of stoichiometrically-balanced chemical reactions. ChemAlgebra requires manipulating sets of complex discrete objects -- molecules represented as formulas or graphs -- under algebraic constraints such as the mass preservation principle. We believe that ChemAlgebra can serve as a useful test bed for the next generation of machine reasoning models and as a promoter of their development.

Andrea Valenti, Davide Bacciu

The recently introduced weakly disentangled representations proposed to relax some constraints of the previous definitions of disentanglement, in exchange for more flexibility. However, at the moment, weak disentanglement can only be achieved by increasing the amount of supervision as the number of factors of variations of the data increase. In this paper, we introduce modular representations for weak disentanglement, a novel method that allows to keep the amount of supervised information constant with respect the number of generative factors. The experiments shows that models using modular representations can increase their performance with respect to previous work without the need of additional supervision.

Stefano Carotti, Marco Pacini, Alessio Gravina et al.

Graph Neural Networks (GNNs) and Graph Transformers (GTs) are now a fundamental paradigm for graph learning, combining the representation-learning capabilities of deep models with the sample efficiency induced by their inductive biases. Despite their effectiveness, a large body of work has shown that these models still face fundamental limitations in tasks that require capturing correlations between distant regions of a graph. To address this issue, we introduce Graph Hierarchical Recurrence (GHR), a novel framework that operates jointly on the input graph and on a hierarchical abstraction obtained through pooling. We also show that the limitations of existing models are even more pronounced in out-of-range generalization, where test instances involve interactions over distances longer than those observed during training. By contrast, despite its simple design, GHR provides three key advantages: strong performance on long-range dependencies, improved out-of-range generalization, and high parameter efficiency. To corroborate these claims, we show that across a broad set of long-range benchmarks, GHR consistently outperforms existing graph models while using as little as 1% of the parameters of current state-of-the-art models. These results suggest a complementary direction to the current trend of scaling architectures to obtain graph foundation models, indicating that increased model capacity alone may not be sufficient for generalization.

Andrea Cossu, Federico Giannini, Giacomo Ziffer et al.

Continual Learning (CL) and Streaming Machine Learning (SML) study the ability of agents to learn from a stream of non-stationary data. Despite sharing some similarities, they address different and complementary challenges. While SML focuses on rapid adaptation after changes (concept drifts), CL aims to retain past knowledge when learning new tasks. After a brief introduction to CL and SML, we discuss Streaming Continual Learning (SCL), an emerging paradigm providing a unifying solution to real-world problems, which may require both SML and CL abilities. We claim that SCL can i) connect the CL and SML communities, motivating their work towards the same goal, and ii) foster the design of hybrid approaches that can quickly adapt to new information (as in SML) without forgetting previous knowledge (as in CL). We conclude the paper with a motivating example and a set of experiments, highlighting the need for SCL by showing how CL and SML alone struggle in achieving rapid adaptation and knowledge retention.

Danilo Numeroso, Davide Bacciu, Petar Veličković

Path finding in graphs is one of the most studied classes of problems in computer science. In this context, search algorithms are often extended with heuristics for a more efficient search of target nodes. In this work we combine recent advancements in Neural Algorithmic Reasoning to learn efficient heuristic functions for path finding problems on graphs. At training time, we exploit multi-task learning to learn jointly the Dijkstra's algorithm and a consistent heuristic function for the A* search algorithm. At inference time, we plug our learnt heuristics into the A* algorithm. Results show that running A* over the learnt heuristics value can greatly speed up target node searching compared to Dijkstra, while still finding minimal-cost paths.

Riccardo Massidda, Francesco Landolfi, Martina Cinquini et al.

The structure learning problem consists of fitting data generated by a Directed Acyclic Graph (DAG) to correctly reconstruct its arcs. In this context, differentiable approaches constrain or regularize the optimization problem using a continuous relaxation of the acyclicity property. The computational cost of evaluating graph acyclicity is cubic on the number of nodes and significantly affects scalability. In this paper we introduce COSMO, a constraint-free continuous optimization scheme for acyclic structure learning. At the core of our method, we define a differentiable approximation of an orientation matrix parameterized by a single priority vector. Differently from previous work, our parameterization fits a smooth orientation matrix and the resulting acyclic adjacency matrix without evaluating acyclicity at any step. Despite the absence of explicit constraints, we prove that COSMO always converges to an acyclic solution. In addition to being asymptotically faster, our empirical analysis highlights how COSMO performance on graph reconstruction compares favorably with competing structure learning methods.

Luca Miglior, Matteo Tolloso, Alessio Gravina et al.

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.

Tai Hoang, Alessandro Trenta, Alessio Gravina et al.

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 to reduce 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.

Leonardo Stoppani, Davide Bacciu, Shahab Mokarizadeh

Diffusion models have become widely popular for automated floorplan generation, producing highly realistic layouts conditioned on user-defined constraints. However, optimizing for perceptual metrics such as the Fréchet Inception Distance (FID) causes limited design diversity. To address this, we propose the Diversity Score (DS), a metric that quantifies layout diversity under fixed constraints. Moreover, to improve geometric consistency, we introduce a Boundary Cross-Attention (BCA) module that enables conditioning on building boundaries. Our experiments show that BCA significantly improves boundary adherence, while prolonged training drives diversity collapse undiagnosed by FID, revealing a critical trade-off between realism and diversity. Out-Of-Distribution evaluations further demonstrate the models' reliance on dataset priors, emphasizing the need for generative systems that explicitly balance fidelity, diversity, and generalization in architectural design tasks.

Tommy Woodley, Shireen Kudukkil Manchingal, Matteo Tolloso et al.

Uncertainty quantification has become an important factor in understanding the data representations produced by Graph Neural Networks (GNNs). Despite their predictive capabilities being ever useful across industrial workspaces, the inherent uncertainty induced by the nature of the data is a huge mitigating factor to GNN performance. While aleatoric uncertainty is the result of noisy and incomplete stochastic data such as missing edges or over-smoothing, epistemic uncertainty arises from lack of knowledge about a system or model (e.g., a graph's topology or node feature representation), which can be reduced by gathering more data and information. In this paper, we propose an original new framework in which node-level epistemic uncertainty is modelled in a belief function (finite random set) formalism. The resulting Random-Set Graph Neural Networks have a belief-function head predicting a random set over the list of classes, from which both a precise probability prediction and a measure of epistemic uncertainty can be obtained. Extensive experiments on 9 different graph learning datasets, including real-world autonomous driving benchmarks as such Nuscene and ROAD, demonstrate RS-GNN's superior uncertainty quantification capabilities

Michele Resta, Davide Bacciu

Modern Neural Machine Translation systems exhibit strong performance in several different languages and are constantly improving. Their ability to learn continuously is, however, still severely limited by the catastrophic forgetting issue. In this work, we leverage a key property of encoder-decoder Transformers, i.e. their generative ability, to propose a novel approach to continually learning Neural Machine Translation systems. We show how this can effectively learn on a stream of experiences comprising different languages, by leveraging a replay memory populated by using the model itself as a generator of parallel sentences. We empirically demonstrate that our approach can counteract catastrophic forgetting without requiring explicit memorization of training data. Code will be publicly available upon publication. Code: https://github.com/m-resta/sg-rep

Giacomo Lanciano, Filippo Galli, Tommaso Cucinotta et al.

Cloud auto-scaling mechanisms are typically based on reactive automation rules that scale a cluster whenever some metric, e.g., the average CPU usage among instances, exceeds a predefined threshold. Tuning these rules becomes particularly cumbersome when scaling-up a cluster involves non-negligible times to bootstrap new instances, as it happens frequently in production cloud services. To deal with this problem, we propose an architecture for auto-scaling cloud services based on the status in which the system is expected to evolve in the near future. Our approach leverages on time-series forecasting techniques, like those based on machine learning and artificial neural networks, to predict the future dynamics of key metrics, e.g., resource consumption metrics, and apply a threshold-based scaling policy on them. The result is a predictive automation policy that is able, for instance, to automatically anticipate peaks in the load of a cloud application and trigger ahead of time appropriate scaling actions to accommodate the expected increase in traffic. We prototyped our approach as an open-source OpenStack component, which relies on, and extends, the monitoring capabilities offered by Monasca, resulting in the addition of predictive metrics that can be leveraged by orchestration components like Heat or Senlin. We show experimental results using a recurrent neural network and a multi-layer perceptron as predictor, which are compared with a simple linear regression and a traditional non-predictive auto-scaling policy. However, the proposed framework allows for the easy customization of the prediction policy as needed.

Vincenzo Lomonaco, Lorenzo Pellegrini, Andrea Cossu et al.

Learning continually from non-stationary data streams is a long-standing goal and a challenging problem in machine learning. Recently, we have witnessed a renewed and fast-growing interest in continual learning, especially within the deep learning community. However, algorithmic solutions are often difficult to re-implement, evaluate and port across different settings, where even results on standard benchmarks are hard to reproduce. In this work, we propose Avalanche, an open-source end-to-end library for continual learning research based on PyTorch. Avalanche is designed to provide a shared and collaborative codebase for fast prototyping, training, and reproducible evaluation of continual learning algorithms.

Alessio Gravina, Moshe Eliasof, Claudio Gallicchio et al.

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.

Alessio Gravina, Daniele Zambon, Davide Bacciu et al.

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.

Marco Lepri, Davide Bacciu, Cosimo Della Santina

This work concerns control-oriented and structure-preserving learning of low-dimensional approximations of high-dimensional physical systems, with a focus on mechanical systems. We investigate the integration of neural autoencoders in model order reduction, while at the same time preserving Hamiltonian or Lagrangian structures. We focus on extensively evaluating the considered methodology by performing simulation and control experiments on large mass-spring-damper networks, with hundreds of states. The empirical findings reveal that compressed latent dynamics with less than 5 degrees of freedom can accurately reconstruct the original systems' transient and steady-state behavior with a relative total error of around 4\%, while simultaneously accurately reconstructing the total energy. Leveraging this system compression technique, we introduce a model-based controller that exploits the mathematical structure of the compressed model to regulate the configuration of heavily underactuated mechanical systems.

Matteo Ninniri, Marco Podda, Davide Bacciu

This work focuses on the task of property targeting: that is, generating molecules conditioned on target chemical properties to expedite candidate screening for novel drug and materials development. DiGress is a recent diffusion model for molecular graphs whose distinctive feature is allowing property targeting through classifier-based (CB) guidance. While CB guidance may work to generate molecular-like graphs, we hint at the fact that its assumptions apply poorly to the chemical domain. Based on this insight we propose a classifier-free DiGress (FreeGress), which works by directly injecting the conditioning information into the training process. CF guidance is convenient given its less stringent assumptions and since it does not require to train an auxiliary property regressor, thus halving the number of trainable parameters in the model. We empirically show that our model yields up to 79% improvement in Mean Absolute Error with respect to DiGress on property targeting tasks on QM9 and ZINC-250k benchmarks. As an additional contribution, we propose a simple yet powerful approach to improve chemical validity of generated samples, based on the observation that certain chemical properties such as molecular weight correlate with the number of atoms in molecules.

Rudy Semola, Julio Hurtado, Vincenzo Lomonaco et al.

Hyperparameter selection in continual learning scenarios is a challenging and underexplored aspect, especially in practical non-stationary environments. Traditional approaches, such as grid searches with held-out validation data from all tasks, are unrealistic for building accurate lifelong learning systems. This paper aims to explore the role of hyperparameter selection in continual learning and the necessity of continually and automatically tuning them according to the complexity of the task at hand. Hence, we propose leveraging the nature of sequence task learning to improve Hyperparameter Optimization efficiency. By using the functional analysis of variance-based techniques, we identify the most crucial hyperparameters that have an impact on performance. We demonstrate empirically that this approach, agnostic to continual scenarios and strategies, allows us to speed up hyperparameters optimization continually across tasks and exhibit robustness even in the face of varying sequential task orders. We believe that our findings can contribute to the advancement of continual learning methodologies towards more efficient, robust and adaptable models for real-world applications.

Matteo Tolloso, Davide Bacciu

Uncertainty quantification is essential for deploying reliable Graph Neural Networks (GNNs), where existing approaches primarily rely on Bayesian inference or ensembles. In this paper, we introduce the first credal graph neural networks (CGNNs), which extend credal learning to the graph domain by training GNNs to output set-valued predictions in the form of credal sets. To account for the distinctive nature of message passing in GNNs, we develop a complementary approach to credal learning that leverages different aspects of layer-wise information propagation. We assess our approach on uncertainty quantification in node classification under out-of-distribution conditions. Our analysis highlights the critical role of the graph homophily assumption in shaping the effectiveness of uncertainty estimates. Extensive experiments demonstrate that CGNNs deliver more reliable representations of epistemic uncertainty and achieve state-of-the-art performance under distributional shift on heterophilic graphs.

Ali Hariri, Álvaro Arroyo, Alessio Gravina et al.

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.

Lanpei Li, Elia Piccoli, Andrea Cossu et al.

Continual Learning (CL) focuses on maximizing the predictive performance of a model across a non-stationary stream of data. Unfortunately, CL models tend to forget previous knowledge, thus often underperforming when compared with an offline model trained jointly on the entire data stream. Given that any CL model will eventually make mistakes, it is of crucial importance to build calibrated CL models: models that can reliably tell their confidence when making a prediction. Model calibration is an active research topic in machine learning, yet to be properly investigated in CL. We provide the first empirical study of the behavior of calibration approaches in CL, showing that CL strategies do not inherently learn calibrated models. To mitigate this issue, we design a continual calibration approach that improves the performance of post-processing calibration methods over a wide range of different benchmarks and CL strategies. CL does not necessarily need perfect predictive models, but rather it can benefit from reliable predictive models. We believe our study on continual calibration represents a first step towards this direction.

Andrea Pugnana, Riccardo Massidda, Francesco Giannini et al.

Concept Bottleneck Models (CBMs) are machine learning models that improve interpretability by grounding their predictions on human-understandable concepts, allowing for targeted interventions in their decision-making process. However, when intervened on, CBMs assume the availability of humans that can identify the need to intervene and always provide correct interventions. Both assumptions are unrealistic and impractical, considering labor costs and human error-proneness. In contrast, Learning to Defer (L2D) extends supervised learning by allowing machine learning models to identify cases where a human is more likely to be correct than the model, thus leading to deferring systems with improved performance. In this work, we gain inspiration from L2D and propose Deferring CBMs (DCBMs), a novel framework that allows CBMs to learn when an intervention is needed. To this end, we model DCBMs as a composition of deferring systems and derive a consistent L2D loss to train them. Moreover, by relying on a CBM architecture, DCBMs can explain why defer occurs on the final task. Our results show that DCBMs achieve high predictive performance and interpretability at the cost of deferring more to humans.

Elia Piccoli, Malio Li, Giacomo Carfì et al.

The recent focus and release of pre-trained models have been a key components to several advancements in many fields (e.g. Natural Language Processing and Computer Vision), as a matter of fact, pre-trained models learn disparate latent embeddings sharing insightful representations. On the other hand, Reinforcement Learning (RL) focuses on maximizing the cumulative reward obtained via agent's interaction with the environment. RL agents do not have any prior knowledge about the world, and they either learn from scratch an end-to-end mapping between the observation and action spaces or, in more recent works, are paired with monolithic and computationally expensive Foundational Models. How to effectively combine and leverage the hidden information of different pre-trained models simultaneously in RL is still an open and understudied question. In this work, we propose Weight Sharing Attention (WSA), a new architecture to combine embeddings of multiple pre-trained models to shape an enriched state representation, balancing the tradeoff between efficiency and performance. We run an extensive comparison between several combination modes showing that WSA obtains comparable performance on multiple Atari games compared to end-to-end models. Furthermore, we study the generalization capabilities of this approach and analyze how scaling the number of models influences agents' performance during and after training.

Andrea Ceni, Alessio Gravina, Claudio Gallicchio et al.

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.

Luca Miglior, Lorenzo Simone, Marco Podda et al.

Optimizing chemical properties is a challenging task due to the vastness and complexity of chemical space. Here, we present a generative energy-based architecture for implicit chemical property optimization, designed to efficiently generate molecules that satisfy target properties without explicit conditional generation. We use Graph Energy Based Models and a training approach that does not require property labels. We validated our approach on well-established chemical benchmarks, showing superior results to state-of-the-art methods and demonstrating robustness and efficiency towards de novo drug design.

Moshe Eliasof, Alessio Gravina, Andrea Ceni et al.

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 Autoregressive Moving Average (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. Extensive experiments on 14 synthetic and real-world datasets demonstrate that GRAMA consistently outperforms backbone models and performs competitively with state-of-the-art methods.

Asma Sattar, Georgios Deligiorgis, Marco Trincavelli et al.

Dynamic multi-relational graphs are an expressive relational representation for data enclosing entities and relations of different types, and where relationships are allowed to vary in time. Addressing predictive tasks over such data requires the ability to find structure embeddings that capture the diversity of the relationships involved, as well as their dynamic evolution. In this work, we establish a novel class of challenging tasks for dynamic multi-relational graphs involving out-of-domain link prediction, where the relationship being predicted is not available in the input graph. We then introduce a novel Graph Neural Network model, named GOOD, designed specifically to tackle the out-of-domain generalization problem. GOOD introduces a novel design concept for multi-relation embedding aggregation, based on the idea that good representations are such when it is possible to disentangle the mixing proportions of the different relational embeddings that have produced it. We also propose five benchmarks based on two retail domains, where we show that GOOD can effectively generalize predictions out of known relationship types and achieve state-of-the-art results. Most importantly, we provide insights into problems where out-of-domain prediction might be preferred to an in-domain formulation, that is, where the relationship to be predicted has very few positive examples.

William Leeney, Alessio Gravina, Davide Bacciu

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.