Identifying and interpreting tuning dimensions in deep networks
Preprint
Abstract
In neuroscience, a tuning dimension is a stimulus attribute that accounts for much of the activation variance of a group of neurons. These are commonly used to decipher the responses of such groups. While researchers have attempted to manually identify an analogue to these tuning dimensions in deep neural networks, we are unaware of an automatic way to discover them. This work contributes an unsupervised framework for identifying and interpreting "tuning dimensions" in deep networks. Our method correctly identifies the tuning dimensions of a synthetic Gabor filter bank and tuning dimensions of the first two layers of InceptionV1 trained on ImageNet.Keywords
All Related Versions
This publication has 14 references indexed in Scilit:
- The What-If Tool: Interactive Probing of Machine Learning ModelsIEEE Transactions on Visualization and Computer Graphics, 2019
- Evolving Images for Visual Neurons Using a Deep Generative Network Reveals Coding Principles and Neuronal PreferencesCell, 2019
- Going deeper with convolutions2015
- ImageNet Large Scale Visual Recognition ChallengeInternational Journal of Computer Vision, 2015
- A Logarithmic, Scale-Invariant Representation of Speed in Macaque Middle Temporal Area Accounts for Speed Discrimination PerformanceThe Journal of Neuroscience, 2005
- Nonlinear Dimensionality Reduction by Locally Linear EmbeddingScience, 2000
- Independent component analysis: algorithms and applicationsNeural Networks, 2000
- Emergence of simple-cell receptive field properties by learning a sparse code for natural imagesNature, 1996
- Neuronal Population Coding of Movement DirectionScience, 1986
- Receptive fields, binocular interaction and functional architecture in the cat's visual cortexThe Journal of Physiology, 1962