How deep learning works --The geometry of deep learning

Abstract

Why and how that deep learning works well on different tasks remains a mystery from a theoretical perspective. In this paper we draw a geometric picture of the deep learning system by finding its analogies with two existing geometric structures, the geometry of quantum computations and the geometry of the diffeomorphic template matching. In this framework, we give the geometric structures of different deep learning systems including convolutional neural networks, residual networks, recursive neural networks, recurrent neural networks and the equilibrium prapagation framework. We can also analysis the relationship between the geometrical structures and their performance of different networks in an algorithmic level so that the geometric framework may guide the design of the structures and algorithms of deep learning systems.

16 pages, 13 figures

References in corpus (3)

• Complexity and Shock Wave Geometries
• Quantum Computation as Geometry
• Switchbacks and the Bridge to Nowhere

Cited by in corpus (5)

• Geometrization of deep networks for the interpretability of deep learning systems
• Demystifying AlphaGo Zero as AlphaGo GAN
• Understanding over-parameterized deep networks by geometrization
• Deep network as memory space: complexity, generalization, disentangled representation and interpretability
• Gauge theory and twins paradox of disentangled representations