EMERGE’s findings will be communicated at scientific conferences and published in open-access journals. Find below the current list of publications.
Access all publication on the project's Zenodo community clicking here.
Residual Echo State Networks: Residual recurrent neural networks with stable dynamics and fast learning
This paper studies the architectural bias of residual connections in the context of recurrent neural networks (RNNs), specifically in the temporal dimension.
Andrea Ceni, Claudio Gallicchio, Neurocomputing, 597, 2024, 127966, doi: 10.1016/j.neucom.2024.127966.
Residual Reservoir Computing Neural Networks for Time-series Classification
In this paper, the authors augment standard Echo State Networks (ESNs) with linear reservoir-skip connections modulated by an untrained orthogonal weight matrix.
Andrea Ceni and Claudio Gallicchio, in ESANN 2023 proceedings, European Symposium on Artificial Neural Networks, Computational Intelligence and Machine Learning. Bruges (Belgium), doi: 10.14428/esann/2023.ES2023-112
Residual Reservoir Memory Networks
This paper introduces a novel class of untrained Recurrent Neural Networks (RNNs) within the Reservoir Computing (RC) paradigm, called Residual Reservoir Memory Networks (ResRMNs).
M. Pinna, A. Ceni and C. Gallicchio, 2025 International Joint Conference on Neural Networks (IJCNN), Rome, Italy, 2025, pp. 1-7, doi: 10.1109/IJCNN64981.2025.11227859.
Safe Control of Soft Robots: Bridging Physics and Learned Models
The contributions presented in this abstract equip soft robots with the motor intelligence they need to function effectively in human-centric environments.
Stölzle, M. Abstract from 8th IEEE-RAS International Conference on Soft Robotics (2025), Lausanne, Switzerland.
Shared Awareness Across Domain-Specific Artificial Intelligence: An Alternative to Domain-General Intelligence and Artificial Consciousness
This paper argues that it will not require machines to be conscious and that simpler ways of sharing awareness are sufficient.
Deroy, O., Bacciu, D., Bahrami, B., Della Santina, C. and Hauert, S. (2024), Adv. Intell. Syst., 6: 2300740. doi: 10.1002/aisy.202300740.
Singular-Perturbation Control of a Tendon-Driven Soft Robot: Theory and Experiments
This work analyzes the influence of the actuation dynamics in tendon-driven continuum soft robots performing trajectory-tracking tasks.
L. N. Ribeiro, P. Borja, C. D. Santina and B. Deutschmann, in IEEE Transactions on Control Systems Technology, doi: 10.1109/TCST.2025.3546564.
Smell Driven Navigation for Soft Robotic Arms: Artificial Nose and Control
This work proposes an artificial nose on a soft robotic arm that ensures separate smell concentration readings.
F. Piqué, F. Stella, J. Hughes, E. Falotico and C. D. Santina, 2023 IEEE International Conference on Soft Robotics (RoboSoft), Singapore, Singapore, 2023, pp. 1-7, doi: 10.1109/RoboSoft55895.2023.10122116.
Social Judgment Interferes With Action Generation During Instrumental Learning
This work investigated how social judgment modulates behavioral and neural mechanisms underlying value-based decision-making.
Zahra Rezazadeh, Bahador Bahrami, and Simone Schütz-Bosbach, Journal of Psychophysiology 2025 39:2, 74-86. DOI: 10.1027/0269-8803/a000348
Soft Robot Shape Estimation With IMUs Leveraging PCC Kinematics for Drift Filtering
This letter proposes a method to eliminate this limitation by leveraging the Piecewise Constant Curvature model assumption.
F. Stella, C. D. Santina and J. Hughes, in IEEE Robotics and Automation Letters, vol. 9, no. 2, pp. 1945-1952, Feb. 2024, doi: 10.1109/LRA.2023.3339063.
SONAR: Long-Range Graph Propagation Through Information Waves
This work introduces SONAR, a novel GNN architecture inspired by the dynamics of wave propagation in continuous media.
Alessandro Trenta and Alessio Gravina and Davide Bacciu, in proceedings of The Thirty-ninth Annual Conference on Neural Information Processing Systems, 2025.
Sparse assemblies of recurrent neural networks with stability guarantees
This work introduces AdaDiag, a framework for constructing sparse assemblies of recurrent neural networks (RNNs) with formal stability guarantees.
Andrea Ceni, Valerio De Caro, Davide Bacciu, Claudio Gallicchio, Neurocomputing, v. 675, 2026, p. 132952, DOI: 10.1016/j.neucom.2026.132952.
Sparse Autoencoders Find Partially Interpretable Features in Italian Small Language Models
This work provides an early evaluation on the feasibility of using Sparse Autoencoders to interpret models trained to be natively Italian.
Alessandro Bondielli, Lucia Passaro, and Alessandro Lenci, in CLiC-it 2025: Eleventh Italian Conference on Computational Linguistics, September 24 — 26, 2025, Cagliari, Italy.
Sparse Reservoir Topologies for Physical Implementations of Random Oscillators Networks
In this paper, the authors propose 6 sparse topologies for Random Oscillators Networks (RON) and study the performance of the model across different levels of sparsity and different numbers of hidden units.
Andrea Cossu, Andrea Ceni, Davide Bacciu, and Claudio Gallicchio. 2025. In Proceedings of the 4th International Conference on AI-ML Systems (AIMLSystems '24). DOI: 10.1145/3703412.3703413
SpikingSoft: A Spiking Neuron Controller for Bio-inspired Locomotion with Soft Snake Robots
This work introduces the Double Threshold Spiking neuron model with adjustable thresholds to generate varied output patterns.
C. Zhang, C. Wang, W. Pan and C. D. Santina, 2025 IEEE 8th International Conference on Soft Robotics (RoboSoft), Lausanne, Switzerland, 2025, pp. 1-8, doi: 10.1109/RoboSoft63089.2025.11020907.
Streaming Continual Learning for Unified Adaptive Intelligence in Dynamic Environments
This paper puts forward a unified setting that harnesses the benefits of both Continual learning (CL) and streaming machine learning (SML): their ability to quickly adapt to nonstationary data streams without forgetting previous knowledge.
F. Giannini, G. Ziffer, A. Cossu and V. Lomonaco, in IEEE Intelligent Systems, vol. 39, no. 6, pp. 81-85, Nov.-Dec. 2024, doi: 10.1109/MIS.2024.3479469.
Symbolic learning of interpretable reduced-order models for jumping quadruped robots
This work proposes a methodology that combines a linear autoencoder with symbolic regression to derive reduced-order models that are central to motion planning and control of quadruped robots.
Gioele Buriani, Jingyue Liu, Maximilian Stölzle, Cosimo Della Santina, Jiatao Ding, Symbolic learning of interpretable reduced-order models for jumping quadruped robots, IFAC Journal of Systems and Control, Volume 35, 2026, 100360, DOI: 10.1016/j.ifacsc.2025.100360.
Synergy-based robotic quadruped leveraging passivity for natural intelligence and behavioural diversity
This study shows a bioinspired approach to the design of quadrupeds that seeks to exploit the body and the passive properties of the robot while maintaining active controllability on the system through minimal actuation.
Stella, F., Achkar, M.M., Della Santina, C. et al. Nat Mach Intell 7, 386–399 (2025). DOI: 10.1038/s42256-025-00988-x
Temporal Graph ODEs for Irregularly-Sampled Time Series
This work introduces 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.
Alessio Gravina, Daniele Zambon, Davide Bacciu and Cesare Alippi, in Proceedings of the Thirty-Third International Joint Conference on Artificial Intelligence, IJCAI-24, 4025-4034, doi: 10.24963/ijcai.2024/445.
