Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

The Deep Learning model of Higher-Lower-Order Cognition, Memory, and Affection- More General Than KAN

Jun-Bo Tao, Bai-Qing Sun, Wei-Dong Zhu, Shi-You Qu, Jia-Qiang Li, Guo-Qi Li, Yan-Yan Wang, Ling-Kun Chen, Chong Wu, Yu Xiong, Jiaxuan Zhou

TL;DR

This work introduces the Plasticity Neural Network (PNN), an RNN-based framework augmenting traditional synaptic weights with dynamic synaptic-range weights and memory-gradient information to model brain plasticity and memory consolidation. By incorporating astrocyte-mediated factors and higher/lower-order Taylor approximations, the authors develop variants (CRPNN, RRPNN, ORPNN) and demonstrate that memory-aware forms (ORPNN) better fit empirical data, especially around critical periods. The methodology blends concepts from neuroscience (memory engrams, astrocyte function, synaptic competition) with DL optimization (gradient, Newton updates) and even explores non-classical ideas like turbulence and memory turbulence through a logarithmic-spiral memory framework. The results suggest astrocyte factors and memory gradients improve convergence and predictive accuracy, with potential implications for understanding aging and Alzheimer's disease, and for designing brain-inspired adaptive learning systems. Overall, the paper contributes a comprehensive, biologically grounded DL model that unifies synaptic strength, synaptic-range plasticity, memory dynamics, and higher/lower-order cognition.

Abstract

We firstly simulated disease dynamics by KAN (Kolmogorov-Arnold Networks) nearly 4 years ago, but the kernel functions in the edge include the exponential number of infected and discharged people and is also in line with the Kolmogorov-Arnold representation theorem, and the shared weights in the edge are the infection rate and cure rate, and used activation function by tanh at the node of edge. And this Arxiv preprint version 1 of March 2022 is an upgraded version of KAN, considering the invariant coarse-grained which calculated by residual or gradient of MSE loss. The improved KAN is PNN (Plasticity Neural Networks) or ELKAN (Edge Learning KNN), in addition to edge learning, it also considered the trimming of the edge. We not inspired by the Kolmogorov-Arnold representation theorem but inspired by the brain science. The ELKAN to explain brain, the variables correspond to different types of neurons, the learning edge can be explained by rebalance of synaptic strength and glial cells phagocytose synapses, and the kernel function means the discharge of neurons and synapses, different neurons and edges mean brain regions. Through testing by cosine, the ELKAN or ORPNN (Optimized Range PNN) is better than the KAN or CRPNN (Constant Range PNN).The ELKAN is more general to explore brain, such as mechanism of consciousness, interactions of natural frequencies in brain regions, synaptic and neuronal discharge frequencies, and data signal frequencies; mechanism of Alzheimer's disease, the Alzheimer's patients has more high frequencies in the upstream brain regions; long short-term relatively good and inferior memory which means gradient of architecture and architecture; turbulent energy flow in different brain regions, turbulence critical conditions need to be met; heart-brain of the quantum entanglement may occur between the emotions of heartbeat and the synaptic strength of brain potentials.

The Deep Learning model of Higher-Lower-Order Cognition, Memory, and Affection- More General Than KAN

TL;DR

This work introduces the Plasticity Neural Network (PNN), an RNN-based framework augmenting traditional synaptic weights with dynamic synaptic-range weights and memory-gradient information to model brain plasticity and memory consolidation. By incorporating astrocyte-mediated factors and higher/lower-order Taylor approximations, the authors develop variants (CRPNN, RRPNN, ORPNN) and demonstrate that memory-aware forms (ORPNN) better fit empirical data, especially around critical periods. The methodology blends concepts from neuroscience (memory engrams, astrocyte function, synaptic competition) with DL optimization (gradient, Newton updates) and even explores non-classical ideas like turbulence and memory turbulence through a logarithmic-spiral memory framework. The results suggest astrocyte factors and memory gradients improve convergence and predictive accuracy, with potential implications for understanding aging and Alzheimer's disease, and for designing brain-inspired adaptive learning systems. Overall, the paper contributes a comprehensive, biologically grounded DL model that unifies synaptic strength, synaptic-range plasticity, memory dynamics, and higher/lower-order cognition.

Abstract

We firstly simulated disease dynamics by KAN (Kolmogorov-Arnold Networks) nearly 4 years ago, but the kernel functions in the edge include the exponential number of infected and discharged people and is also in line with the Kolmogorov-Arnold representation theorem, and the shared weights in the edge are the infection rate and cure rate, and used activation function by tanh at the node of edge. And this Arxiv preprint version 1 of March 2022 is an upgraded version of KAN, considering the invariant coarse-grained which calculated by residual or gradient of MSE loss. The improved KAN is PNN (Plasticity Neural Networks) or ELKAN (Edge Learning KNN), in addition to edge learning, it also considered the trimming of the edge. We not inspired by the Kolmogorov-Arnold representation theorem but inspired by the brain science. The ELKAN to explain brain, the variables correspond to different types of neurons, the learning edge can be explained by rebalance of synaptic strength and glial cells phagocytose synapses, and the kernel function means the discharge of neurons and synapses, different neurons and edges mean brain regions. Through testing by cosine, the ELKAN or ORPNN (Optimized Range PNN) is better than the KAN or CRPNN (Constant Range PNN).The ELKAN is more general to explore brain, such as mechanism of consciousness, interactions of natural frequencies in brain regions, synaptic and neuronal discharge frequencies, and data signal frequencies; mechanism of Alzheimer's disease, the Alzheimer's patients has more high frequencies in the upstream brain regions; long short-term relatively good and inferior memory which means gradient of architecture and architecture; turbulent energy flow in different brain regions, turbulence critical conditions need to be met; heart-brain of the quantum entanglement may occur between the emotions of heartbeat and the synaptic strength of brain potentials.
Paper Structure (10 sections, 21 equations, 38 figures, 21 tables)

This paper contains 10 sections, 21 equations, 38 figures, 21 tables.

Table of Contents

  1. Introduction
  2. Method
  3. Discussion
  4. Comparison of PNNs in some scenarios
  5. Synapse formation affects brain function - inhibiting dendrites
  6. Effects of astrocytes are reflected in astrocytes phagocytose synapses factor and cortex memory persistence factor during critical and closure critical period
  7. A simple PNN
  8. The simplified higher-lower-order Deep Learning model
  9. Mechanism, Interpretability, Findings and Hypotheses
  10. Conclusion

Figures (38)

  • Figure 1: The brain's reverse turbulence is explained by figures: (a)Biological experiment results bib33 (b)Normal IQ (c)High IQ (d)Low IQ (e)$LEN$ points of relatively good and inferior memories (f)The critical angle between turbulence and laminar flow bib39 (g) Memory, cognition and affection Loss (h) Deep Learning model for upstream and downstream brain regions (i) Cerebral artery, from Gray's Atlas of Anatomy 3rd Edition (j) Angles of logarithmic Spiral and loss of memory engram
  • Figure 2: Astrocytic synapse formation cortex memory persistence factor and astrocytes phagocytose synapses factor: (a) Astrocytic synapse formation cortex memory persistence factor (b) Astrocytes phagocytose synapses factor
  • Figure 3: CRPNN, RRPNN and ORPNN flow charts.
  • Figure 4: first test results belong to CRPNN, RRPNN, and ORPNN respectively
  • Figure 5: second test results belong to CRPNN, RRPNN, and ORPNN respectively
  • ...and 33 more figures