Fernández Sánchez, AlbertoGestal, M.Dorado, JuliánPazos, A.2026-02-112026-02-112026Fernandez-Sanchez, A., Gestal, M., Dorado, J., Pazos, A. (2026). From Biological Neurons to Artificial Neural Networks: A Bioinspired Training Alternative. In: Rojas, I., Joya, G., Catala, A. (eds) Advances in Computational Intelligence. IWANN 2025. Lecture Notes in Computer Science, vol 16008. Springer, Cham. https://doi-org.accedys.udc.es/10.1007/978-3-032-02725-2_22978-3-032-02725-21611-3349https://hdl.handle.net/2183/47364Presentado en: 18th International Work-Conference on Artificial Neural Networks, IWANN 2025, A Coruña, Spain, June 16–18, 2025 This version of the article has been accepted for publication, after peer review (when applicable) and is subject to Springer Nature’s AM terms of use, but is not the Version of Record and does not reflect post-acceptance improvements, or any corrections. The Version of Record is available online at: https://doi.org/10.1007/978-3-032-02725-2_22[Abstract]: Artificial neural networks often rely on fixed architectures and uniform training strategies, overlooking adaptive mechanisms found in biological learning. This work presents a conceptual framework for a biologically inspired training algorithm that mimics human skill acquisition through a progressive, difficulty-driven curriculum. Grounded in principles such as synaptic plasticity, dendritic computation, and modular learning, the proposed method restructures training into stages of increasing complexity, expanding the model’s architecture over time. Unlike other biologically inspired approaches that redesign neuron structures, this strategy retains standard neural components, allowing its application to conventional neural architectures while introducing an adaptive training schedule aligned with neurodevelopment. The model incrementally grows to match input difficulty, transferring and perturbing learned weights across stages to promote generalization and efficiency. Although not yet empirically validated, this work outlines a plan for systematic evaluation. Future phases will include the implementation of the training methods and generation of staged models. The framework offers a scalable and interpretable path toward energy-efficient learning and contributes to bridging biological and artificial intelligence.eng© 2026 The Author(s), under exclusive license to Springer Nature Switzerland AGBiologically inspired learningCurriculum learningDendritic computationAdaptive architecturesGreen machine learningconference outputembargoed access10.1007/978-3-032-02725-2_22