The Sparse Manifold Transform

2018/06/23 by Yubei Chen, Dylan M. Paiton, Chen, Yubei +3 · 1 voice · 4 citations
#stat.ML #cs.LG #eess.IV

paper · pdf · doi:10.48550/arxiv.1806.08887

Abstract

We present a signal representation framework called the sparse manifold transform that combines key ideas from sparse coding, manifold learning, and slow feature analysis. It turns non-linear transformations in the primary sensory signal space into linear interpolations in a representational embedding space while maintaining approximate invertibility. The sparse manifold transform is an unsupervised and generative framework that explicitly and simultaneously models the sparse discreteness and low-dimensional manifold structure found in natural scenes. When stacked, it also models hierarchical composition. We provide a theoretical description of the transform and demonstrate properties of the learned representation on both synthetic data and natural videos.

Cited by

• Toward a mechanistic understanding of inference in visual cortex and diffusion models
• Weaver: Kronecker Product Approximations of Spatiotemporal Attention for Traffic Network Forecasting
• Cloning Deterministic Worlds: The Critical Role of Latent Geometry in Long-Horizon World Models
• Scaling Non-Parametric Sampling with Representation

Discussions

• The Sparse Manifold Transform (2018) [pdf] [hn, 36 points, 1 comments]

Related

• Predictive Sparse Manifold Transform
• Learning Signal-Agnostic Manifolds of Neural Fields
• Transformational Sparse Coding
• Learning Efficient Coding of Natural Images with Maximum Manifold Capacity Representations
• Statistical Mechanics of Neural Processing of Object Manifolds