An Unsupervised Algorithm For Learning Lie Group Transformations
arXiv:1001.1027
Abstract
We present several theoretical contributions which allow Lie groups to be fit to high dimensional datasets. Transformation operators are represented in their eigen-basis, reducing the computational complexity of parameter estimation to that of training a linear transformation model. A transformation specific "blurring" operator is introduced that allows inference to escape local minima via a smoothing of the transformation space. A penalty on traversed manifold distance is added which encourages the discovery of sparse, minimal distance, transformations between states. Both learning and inference are demonstrated using these methods for the full set of affine transformations on natural image patches. Transformation operators are then trained on natural video sequences. It is shown that the learned video transformations provide a better description of inter-frame differences than the standard motion model based on rigid translation.
Cited by in corpus (13)
- A Probabilistic Theory of Deep Learning
- Learning the Irreducible Representations of Commutative Lie Groups
- Automatic Symmetry Discovery with Lie Algebra Convolutional Network
- Towards Nonlinear Disentanglement in Natural Data with Temporal Sparse Coding
- Addressing the Topological Defects of Disentanglement via Distributed Operators
- Disentangling by Subspace Diffusion
- Learning image transformations without training examples
- Unsupervised Transformation Learning via Convex Relaxations
- Disentangling images with Lie group transformations and sparse coding
- Learning Identity-Preserving Transformations on Data Manifolds
- Deep Autoencoders: From Understanding to Generalization Guarantees
- Joint Estimation of Image Representations and their Lie Invariants
- Transformational Sparse Coding