1 July 2025, ROME, ITALY
Paper submission deadline: January, 30th 2025 (extended)
Decision notification: March, 31st 2025
Reservoir Computing (RC) has emerged as a computationally efficient approach to training recurrent neural networks (RNNs), offering unique capabilities in handling temporal data without the high resource demands of traditional deep learning models. With the increasing need for low-power, fast, and adaptable neural networks across domains like real-time signal processing, edge computing, and neuromorphic systems, RC presents a compelling solution. Yet, as deep learning methods evolve, RC faces critical challenges that call for new theoretical insights, model innovations, and hardware implementations. Specifically, RC must address limitations in scalability, robustness, and adaptability to maintain its relevance alongside modern neural architectures.
This special session aims to advance Reservoir Computing (RC) by tackling key challenges that span theoretical, model, application, and hardware dimensions. A primary goal is to expand RC’s theoretical foundations to enhance stability and adaptability, especially in dynamic environments, drawing on insights from dynamical systems and control theory to make RC more robust and reliable in real-world applications.
The session also seeks to explore hybrid architectures that combine RC with deep learning approaches, such as graph-based and hierarchical models, enabling RC to scale and handle complex, high-dimensional spatiotemporal tasks, including video and text analysis. Another objective is to foster innovative applications in fields requiring efficient, low-latency data processing, such as robotics, neuroscience, and IoT, where RC’s computational efficiency proves advantageous for real-time and resource-constrained contexts.
Moreover, the session promotes advances in hardware implementations to enable energy-efficient AI, supporting RC’s deployment on neuromorphic platforms like photonic, and spintronic substrates. This direction highlights RC’s potential for ultra-fast, low-energy applications critical to embedded systems and edge AI.
Finally, the session aims to foster interdisciplinary collaboration by joining insights from machine learning, physics, neuroscience, and hardware engineering. This collaborative approach will drive innovation, positioning RC as a foundational framework within modern AI that addresses both theoretical challenges and practical needs across various domains.
We invite researchers to submit papers on all aspects of RC research, targeting contributions on theory, models, and applications.
A list of topics relevant to this session includes, but is not limited to, the following:
New Reservoir Computing models and architectures, including Echo State Networks and Liquid State Machines
Hardware, physical and neuromorphic implementations of Reservoir Computing systems
Learning algorithms in Reservoir Computing
Reservoir Computing in Computational Neuroscience
Reservoir Computing on the edge systems
Novel learning algorithms rooted in Reservoir Computing concepts
Novel applications of Reservoir Computing, e.g., to images, video and structured data
Federated and Continual Learning in Reservoir Computing
Deep and modular Reservoir Computing neural networks
Theory of complex and dynamical systems in Reservoir Computing
Extensions of the Reservoir Computing framework, such as Conceptors
The review process for IJCNN 2025 will be double-blind, i.e. reviewers will not know the author's identity (and vice versa). Authors should ensure their anonymity in the submitted papers. In brief: authors' names should not be included in the submitted pdf; please refer to your prior work in the third person wherever possible; a reviewer may be able to deduce the authors' identities by using external resources, such as technical reports published on the web. The availability of information on the web that may allow reviewers to infer the authors' identities does not constitute a breach of the double-blind submission policy.
Full authors' instructions can be found at the following link: https://2025.ijcnn.org/authors/initial-author-instructions
IJCNN 2025 adopts Microsoft CMT as submission system, available at the following link: https://cmt3.research.microsoft.com/IJCNN2025/
To submit to this special session, create a new submission (main track) and select "Reservoir computing in the deep learning era: theory, models, applications, and hardware implementations" as the Primary Subject area (it's on the right hand side of the list).
Claudio Gallicchio (University of Pisa, Italy), Andrea Ceni (University of Pisa, Italy), Gouhei Tanaka (Nagoya Institute of Technology, Japan), Ziqiang Li (The University of Tokyo, Japan), Xavier Hinaut (Institut National de Recherche en Informatique et en Automatique, France)