Exploratory Analysis of Intrinsic Dimensions
- 0 Collaborators
Intrinsic Dimensions is a fundamental property of data sets in Deep neural networks and help us qualitatively assess its complexity in spatiotemporal space. The aim of the project is to explore the space for interesting results, measure the effect of adversarial attacks, explore its usage for compression, scale to more reinforcement learning problems and establish benchmarks. ...learn more
Project status: Concept
Overview / Usage
The aim of the project is to explore the space for interesting results, measure the effect of adversarial attacks, explore its usage for compression, scale to more reinforcement learning problems and establish benchmarks.
Its usage may help us eliminate a redundant or multidimensional hyperplane ranging the entire solution space and make inferencing cheaper.
Methodology / Approach
Introductory Blog post: https://medium.com/@datafineass/intrinsic-dimensionality-of-a-data-set-pt-1-an-introduction-b7cf2382fe5
Watch out for this space!