The concept of high-order neural networks traces back to the 1970s, where polynomial neural networks were utilized as feature extractors, showcasing their longstanding importance in data handling and machine learning.
Recent advancements in deep learning have paved the way for a renewed integration of polynomial operators into neural networks, enhancing their feature extraction capabilities and opening new avenues for research.
Methodologies for incorporating polynomials in neural networks fall into two categories: establishing polynomial structures and introducing polynomial neurons, which signifies a critical evolution in neural network design.
High-order polynomial networks like the pi-sigma network have been foundational in the evolution of more complex models, influencing how neural networks process and learn from data.
#high-order-neural-networks #polynomial-operators #feature-extraction #deep-learning #neural-networks
Collection
[
|
...
]