2025 Journal article Open Access

Increasing biases can be more efficient than increasing weights
Metta C., Fantozzi M., Papini A., Amato G., Bergamaschi M., Fois A., Galfrè S. G., Marchetti A., Vegliò M., Parton M., Morandin F.
We introduce a novel computational unit for neural networks that features multiple biases, challenging the traditional perceptron structure. This unit emphasizes the importance of preserving uncorrupted information as it is passed from one unit to the next, applying activation functions later in the process with specialized biases for each unit. Through both empirical and theoretical analyses, we show that by focusing on increasing biases rather than weights, there is potential for significant enhancement in a neural network model’s performance. This approach offers an alternative perspective on optimizing information flow within neural networks. See source code (CurioSAI in Increasing biases can be more efficient than increasing weights, 2023. https://github.com/CuriosAI/dac-dev).Source: ADVANCES IN DATA ANALYSIS AND CLASSIFICATION, vol. 119, pp. 437-468
DOI: 10.1007/s11634-025-00649-2
Project(s): SoBigData-PlusPlus
Metrics:

See at: CNR IRIS | link.springer.com | Software Heritage | Advances in Data Analysis and Classification | GitHub | ARUdA | CNR IRIS | CNR IRIS

Improving performance in neural networks by dendrite-activated connection
Metta C., Fantozzi M., Papini A., Amato G., Bergamaschi M., Fois A., Galfre’ S. G., Marchetti A., Vegliò M., Parton M., Morandin F.
We introduce a novel computational unit for neural networks featuring multiple biases, challenging the conventional perceptron structure. Designed to emphasize preserving uncorrupted information as it transfers from one unit to the next, this unit applies activation functions later in the process, incorporating specialized biases for each unit. We posit this unit as an improved design for neural networks and support this with (1) empirical evidence across diverse datasets; (2) a class of functions where this unit utilizes parameters more efficiently; and (3) biological analogies suggesting closer mimicry to natural neural processing. Source code is available at https://github.com/CuriosAI/dac-dev.Source: STUDIES IN CLASSIFICATION, DATA ANALYSIS, AND KNOWLEDGE ORGANIZATION, pp. 133-141
DOI: 10.1007/978-3-031-84702-8_15
Project(s): SoBigData-PlusPlus
Metrics:

See at: CNR IRIS | CNR IRIS | CNR IRIS | link.springer.com

2025 Conference article Open Access

Interactive visual exploration of latent spaces for explainable ai: bridging concepts and features
Metta C., Cappuccio E., Rinzivillo S.
Latent space exploration offers a powerful lens for interpreting and improving the explainability of black-box AI models. This paper introduces a visual interface based on (beta)-Variational Autoencoders that enables users to navigate latent spaces interactively. By employing tools like visual latent sliders and transformation pathways, the interface demonstrates how latent dimensions can influence image representations, uncovering semantic structure and disentangled features. Using the MedMNIST medical image dataset, we illustrate the potential of this approach to bridge the gap between technical latent space analysis and intuitive understanding. Although the focus is on presenting the methodology, this work sets the stage for integrating user interaction and metrics, particularly in high-stakes domains such as medical imaging.Source: CEUR WORKSHOP PROCEEDINGS, vol. 3957, pp. 84-94. Cagliari, Italy, 24/03/2025
Project(s): SoBigData-PlusPlus

See at: ceur-ws.org | CNR IRIS | CNR IRIS

Exploration and generalization in deep learning with SwitchPath activations
Di Cecco A., Papini A., Metta C., Fantozzi M., Galfrè S. G., Morandin F., Parton M.
This work provides a comprehensive theoretical and empirical analysis of SwitchPath, a stochastic activation function that improves learning dynamics by probabilistically toggling between a neuron standard activation and its negation. We develop theoretical foundations and demonstrate its impact in multiple scenarios. By maintaining gradient flow and injecting controlled stochasticity, the method improves generalization, uncertainty estimation, and training efficiency. Experiments in classification show consistent gains over ReLU and Leaky ReLU across CNNs and Vision Transformers, with reduced overfitting and better test accuracy. In generative modeling, a novel two-phase training scheme significantly mitigates mode collapse and accelerates convergence. Our theoretical analysis reveals that SwitchPath introduces a form of multiplicative noise that acts as a structural regularizer. Additional empirical investigations show improved information propagation and reduced model complexity. These results establish this activation mechanism as a simple yet effective way to enhance exploration, regularization, and reliability in modern neural networks.Source: MACHINE LEARNING, vol. 114 (issue 200)
DOI: 10.1007/s10994-025-06840-y
Project(s): SoBigData-PlusPlus
Metrics:

See at: CNR IRIS | CNR IRIS | CNR IRIS | link.springer.com

2025 Conference article Open Access

SwitchPath enhancing exploration in neural networks learning dynamics
Di Cecco A., Papini A., Metta C., Fantozzi M., Galfré S. G., Morandin F., Parton M.
We introduce SwitchPath, a novel stochastic activation function that enhances neural network exploration, performance, and generalization, by probabilistically toggling between the activation of a neuron and its negation. SwitchPath draws inspiration from the analogies between neural networks and decision trees, and from the exploratory and regularizing properties of DropOut as well. Unlike Dropout, which intermittently reduces network capacity by deactivating neurons, Switch- Path maintains continuous activation, allowing networks to dynamically explore alternative information pathways while fully utilizing their capacity. Building on the concept of ε-greedy algorithms to balance exploration and exploitation, SwitchPath enhances generalization capabilities over traditional activation functions. The exploration of alternative paths happens during training without sacrificing computational efficiency. This paper presents the theoretical motivations, practical implementations, and empirical results, showcasing all the described advantages of SwitchPath over established stochastic activation mechanisms.Source: LECTURE NOTES IN COMPUTER SCIENCE, vol. 15243 - Proceedings, Part I, pp. 275-291. Pisa, Italy, 14-16/10/2024
DOI: 10.1007/978-3-031-78977-9_18
Project(s): SoBigData-PlusPlus
Metrics:

See at: CNR IRIS | link.springer.com | CNR IRIS | CNR IRIS

2025 Conference article Open Access

Explainable AI in time-sensitive scenarios prefetched offline explanation model
Russo F. M., Metta C., Monreale A., Rinzivillo S., Pinelli F.
As predictive machine learning models become increasingly adopted and advanced, their role has evolved from merely predicting outcomes to actively shaping them. This evolution has underscored the importance of Trustworthy AI, highlighting the necessity to extend our focus beyond mere accuracy and toward a comprehensive understanding of these models’ behaviors within the specific contexts of their applications. To further progress in explainability, we introduce poem, Prefetched Offline Explanation Model, a model-agnostic, local explainability algorithm for image data. The algorithm generates exemplars, counterexemplars and saliency maps to provide quick and effective explanations suitable for time-sensitive scenarios. Leveraging an existing local algorithm, poem infers factual and counterfactual rules from data to create illustrative examples and opposite scenarios with an enhanced stability by design. A novel mechanism then matches incoming test points with an explanation base and produces diverse exemplars, informative saliency maps and believable counterexemplars. Experimental results indicate that poem outperforms its predecessor abele in speed and ability to generate more nuanced and varied exemplars alongside more insightful saliency maps and valuable counterexemplars.Source: LECTURE NOTES IN COMPUTER SCIENCE, vol. 15244 - Proceedings, Part II, pp. 167-182. Pisa, Italy, 14-16/10/2024
DOI: 10.1007/978-3-031-78980-9_11
Project(s): TANGO , XAI , SoBigData-PlusPlus
Metrics:

See at: CNR IRIS | link.springer.com | CNR IRIS | CNR IRIS

2025 Journal article Open Access

Analyzing RL components for Wagner’s framework via Brouwer’s conjecture
Angileri F., Lombardi G., Fois A., Faraone R., Metta C., Salvi M., Bianchi L. A., Fantozzi M., Galfrè S. G., Pavesi D., Parton M., Morandin F.
In this work, we continue the study and systematization started in our previous work (Angileri et al. in: Lecture notes in computer science 15243 LNAI, 2025, pp 325–338. https://doi.org/10.1007/978-3-031-78977-9_21) of Wagner’s Reinforcement Learning framework to investigate graph conjectures. After identifying three main directions that impact the framework’s performance (the environment dynamics, the RL algorithm and the neural network used as a function approximator), we conduct an ablation study to evaluate the effectiveness of each component, analyzing several variations of them. The experiments compare three environment dynamics, implemented as Gym spaces (Linear, Local and Global), two algorithms [PPO and the Cross-Entropy method (Wagner in Constructions in combinatorics via neural networks, 2021, https://arxiv.org/abs/2104.14516)], different neural network structures (Multi-Layer Perceptron and Graph Neural Networks) and reward systems. This study was intended not only to test the framework’s capabilities, but also to identify a configuration of environment, algorithm, and neural network that can be effective when exploring graph spaces, even with a complex target. For this reason, all the experiments were executed on Brouwer’s Conjecture. We also present the data collected with the various trained models, as these interesting configurations can be used in the inference process on the problem. Our analysis shows that a proper calibration of the individual components of the framework can significantly improve its performance, suggesting effective settings for addressing complex problems and contributing to the study of Brouwer’s Conjecture. All the codes and data are open source and available at https://github.com/CuriosAI/graph_conjectures.Source: MACHINE LEARNING, vol. 114 (issue 242)
DOI: 10.1007/s10994-025-06890-2
Project(s): SoBigData-PlusPlus
Metrics:

See at: CNR IRIS | link.springer.com | CNR IRIS

2025 Conference article Open Access

A systematization of the Wagner framework graph theory conjectures and reinforcement learning
Angileri F., Lombardi G., Fois A., Faraone R., Metta C., Salvi M., Bianchi L. A., Fantozzi M., Galfrè S. G., Pavesi D., Parton M., Morandin F.
In 2021, Adam Zsolt Wagner proposed an approach to disprove conjectures in graph theory using Reinforcement Learning (RL). Wagner frames a conjecture as f(G) < 0 for every graph G, for a certain invariant f; one can then play a single-player graph-building game, where at each turn the player decides whether to add an edge or not. The game ends when all edges have been considered, resulting in a certain graph GT , and f(GT ) is the final score of the game; RL is then used to maximize this score. This brilliant idea is as simple as innovative, and it lends itself to systematic generalization. Several different single-player graph-building games can be employed, along with various RL algorithms. Moreover, RL maximizes the cumulative reward, allowing for step-by-step rewards instead of a single final score, provided the final cumulative reward represents the quantity of interest f(GT ). In this paper, we discuss these and various other choices that can be significant in Wagner’s framework. As a contribution to this systematization, we present four distinct single-player graph-building games. Each game employs both a step-by-step reward system and a single final score. We also propose a principled approach to select the most suitable neural network architecture for any given conjecture and introduce a new dataset of graphs labeled with their Laplacian spectra. The games have been implemented as environments in the Gymnasium framework, and along with the dataset and a simple interface to play with the environments, are available at https://github.com/CuriosAI/graph_conjectures.Source: LECTURE NOTES IN COMPUTER SCIENCE, vol. 15243 - Proceedings, Part I, pp. 325-338. Pisa, Italy, 14-16/10/2024
DOI: 10.1007/978-3-031-78977-9_21
Project(s): SoBigData-PlusPlus
Metrics:

See at: CNR IRIS | link.springer.com | CNR IRIS | CNR IRIS