APTx Neuron: A Unified Trainable Neuron Architecture Integrating Activation and Computation

Fuente: arXiv
Saved in:
Bibliographic Details
Main Author: Kumar, Ravin
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866908732300984320
author Kumar, Ravin
author_facet Kumar, Ravin
contents We propose the APTx Neuron, a novel, unified neural computation unit that integrates non-linear activation and linear transformation into a single trainable expression. The APTx Neuron is derived from the APTx activation function, thereby eliminating the need for separate activation layers and making the architecture both optimization-efficient and elegant. The proposed neuron follows the functional form $y = \sum_{i=1}^{n} ((α_i + \tanh(β_i x_i)) \cdot γ_i x_i) + δ$, where all parameters $α_i$, $β_i$, $γ_i$, and $δ$ are trainable. We validate our APTx Neuron-based architecture on the MNIST dataset, achieving up to $96.69\%$ test accuracy within 11 epochs using approximately 332K trainable parameters. The results highlight the superior expressiveness and training efficiency of the APTx Neuron compared to traditional neurons, pointing toward a new paradigm in unified neuron design and the architectures built upon it. Source code is available at https://github.com/mr-ravin/aptx_neuron.
format Preprint
id arxiv_https___arxiv_org_abs_2507_14270
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle APTx Neuron: A Unified Trainable Neuron Architecture Integrating Activation and Computation
Kumar, Ravin
Neural and Evolutionary Computing
Artificial Intelligence
Computer Vision and Pattern Recognition
Machine Learning
We propose the APTx Neuron, a novel, unified neural computation unit that integrates non-linear activation and linear transformation into a single trainable expression. The APTx Neuron is derived from the APTx activation function, thereby eliminating the need for separate activation layers and making the architecture both optimization-efficient and elegant. The proposed neuron follows the functional form $y = \sum_{i=1}^{n} ((α_i + \tanh(β_i x_i)) \cdot γ_i x_i) + δ$, where all parameters $α_i$, $β_i$, $γ_i$, and $δ$ are trainable. We validate our APTx Neuron-based architecture on the MNIST dataset, achieving up to $96.69\%$ test accuracy within 11 epochs using approximately 332K trainable parameters. The results highlight the superior expressiveness and training efficiency of the APTx Neuron compared to traditional neurons, pointing toward a new paradigm in unified neuron design and the architectures built upon it. Source code is available at https://github.com/mr-ravin/aptx_neuron.
title APTx Neuron: A Unified Trainable Neuron Architecture Integrating Activation and Computation
topic Neural and Evolutionary Computing
Artificial Intelligence
Computer Vision and Pattern Recognition
Machine Learning
url https://arxiv.org/abs/2507.14270