Abstract:Recent progress in artificial intelligence (AI) has been driven by insights from physics and neuroscience, particularly through the development of artificial neural networks (ANNs) capable of complex cognitive tasks such as vision and language processing. Despite these advances, they struggle with continual learning, adaptable knowledge transfer, robustness, and resource efficiency – capabilities that biological systems handle seamlessly. Specifically, neuromorphic systems and artificial neural networks often overlook two key biophysical properties of neural circuits: neuronal diversity and cell-specific neuromodulation. These mechanisms, essential for regulating dynamic learning across brain scales, allow neuromodulators to …
Abstract:Recent progress in artificial intelligence (AI) has been driven by insights from physics and neuroscience, particularly through the development of artificial neural networks (ANNs) capable of complex cognitive tasks such as vision and language processing. Despite these advances, they struggle with continual learning, adaptable knowledge transfer, robustness, and resource efficiency – capabilities that biological systems handle seamlessly. Specifically, neuromorphic systems and artificial neural networks often overlook two key biophysical properties of neural circuits: neuronal diversity and cell-specific neuromodulation. These mechanisms, essential for regulating dynamic learning across brain scales, allow neuromodulators to introduce degeneracy in biological neural networks, ensuring stability and adaptability under changing conditions. In this article, we summarize recent bioinspired models, learning rules, and architectures, and propose a framework for augmenting ANNs, which has the potential to bridge the gap between neuroscience and AI through neurobiological first principles. Our proposed dual-framework approach leverages spiking neural networks to emulate diverse spiking behaviors and dendritic compartmental dynamics, thereby simulating the morphological and functional diversity of neuronal computations. Finally, we outline how integrating these biophysical principles into task-driven spiking neural networks and neuromorphic systems provides scalable solutions for continual learning, adaptability, robustness, and resource-efficiency. Additionally, this approach will not only provide insights into how emergent behaviors arise in neural networks but also catalyze the development of more efficient, reliable, and intelligent neuromorphic systems and robotic agents.
| Comments: | 26 pages, 4 figures, 3 boxes, 1 table |
| Subjects: | Neurons and Cognition (q-bio.NC); Artificial Intelligence (cs.AI); Machine Learning (cs.LG) |
| MSC classes: | 92B20 |
| Cite as: | arXiv:2407.04525 [q-bio.NC] |
| (or arXiv:2407.04525v5 [q-bio.NC] for this version) | |
| https://doi.org/10.48550/arXiv.2407.04525 arXiv-issued DOI via DataCite |
Submission history
From: Alejandro Rodriguez-Garcia [view email] [v1] Fri, 5 Jul 2024 14:11:28 UTC (594 KB) [v2] Mon, 16 Sep 2024 13:49:32 UTC (590 KB) [v3] Tue, 15 Oct 2024 12:51:44 UTC (847 KB) [v4] Mon, 11 Nov 2024 16:58:38 UTC (1,404 KB) [v5] Mon, 3 Nov 2025 13:54:00 UTC (1,832 KB)