Linear Readout of Object Manifolds

Abstract

Objects are represented in sensory systems by continuous manifolds due to sensitivity of neuronal responses to changes in physical features such as location, orientation, and intensity. What makes certain sensory representations better suited for invariant decoding of objects by downstream networks? We present a theory that characterizes the ability of a linear readout network, the perceptron, to classify objects from variable neural responses. We show how the readout perceptron capacity depends on the dimensionality, size, and shape of the object manifolds in its input neural representation.

5 pages, 3 figures, accepted in Physical Review E as Rapid Communication on 14th May. 2016

References in corpus (2)

• Deep Neural Networks Rival the Representation of Primate IT Cortex for Core Visual Object Recognition
• Theory of spike timing based neural classifiers

Cited by in corpus (20)

• Neural population geometry: An approach for understanding biological and artificial neural networks
• Intrinsic dimension of data representations in deep neural networks
• Classification and Geometry of General Perceptual Manifolds
• Dimensionality compression and expansion in Deep Neural Networks
• Statistical learning theory of structured data
• Counting the learnable functions of structured data
• Gaussian Universality of Perceptrons with Random Labels
• Beyond the storage capacity: data driven satisfiability transition
• Linear Classification of Neural Manifolds with Correlated Variability
• On the geometry of generalization and memorization in deep neural networks
• Learning Data Manifolds with a Cutting Plane Method
• Generalization from correlated sets of patterns in the perceptron
• Random features and polynomial rules
• Neural Network Memory Architectures for Autonomous Robot Navigation
• Interpreting Encoding and Decoding Models
• Soft-margin classification of object manifolds
• Statistical Mechanics of Neural Processing of Object Manifolds
• Biologically inspired architectures for sample-efficient deep reinforcement learning
• Simplified derivations for high-dimensional convex learning problems
• Visualizing the geometry of labeled high-dimensional data with spheres