Efficient Learning of Sparse Invariant Representations

2011/05/26 by Karol Gregor, Yann LeCun, Gregor, Karol +1 · 1 citation
Computer Science · #Face and Expression Recognition #Advanced Image and Video Retrieval Techniques #Image Retrieval and Classification Techniques

paper · pdf · doi:10.48550/arxiv.1105.5307

Abstract

We propose a simple and efficient algorithm for learning sparse invariant representations from unlabeled data with fast inference. When trained on short movies sequences, the learned features are selective to a range of orientations and spatial frequencies, but robust to a wide range of positions, similar to complex cells in the primary visual cortex. We give a hierarchical version of the algorithm, and give guarantees of fast convergence under certain conditions.

Cited by

• Metalearning: Sparse Variable-Structure Automata

Related

• Fast Inference in Sparse Coding Algorithms with Applications to Object Recognition
• Trace Quotient with Sparsity Priors for Learning Low Dimensional Image Representations
• Minimizing FLOPs to Learn Efficient Sparse Representations
• Slow Feature Analysis: Unsupervised Learning of Invariances
• Learning Invariant Representations with a Nonparametric Nadaraya-Watson Head