ManifoldFlow: Spectrum-Learnable Stiefel Optimization via SPD Relaxation
cs.LG, stat.ML
Submitted: 2026-07-05
Updated: 2026-10-15
Comments: 39 pages
Code: https://github.com/Hik289/manifold
License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Terminology
Sources
- Orthogonal Weight Normalization: Solution to Optimization over Multiple Dependent Stiefel Manifolds in Deep Neural Networks
- Gated Orthogonal Recurrent Units: On Learning to Forget
- Adam: A Method for Stochastic Optimization
- Geoopt: Riemannian Optimization in PyTorch
- Trivializations for Gradient-Based Optimization on Manifolds
- Intrinsic Muon: Spectral Optimization on Riemannian Matrix Manifolds
- Pointer Sentinel Mixture Models
- All you need is a good init
- Spectral Normalization for Generative Adversarial Networks
- Fashion-MNIST: a Novel Image Dataset for Benchmarking Machine Learning Algorithms
Related papers
- Polynomial-Augmented Neural Networks (PANNs) with Weak Orthogonality Constraints for Enhanced Function and PDE Approximation
- AIRL-S: Unifying Reinforcement Learning and Search-Based Test-Time Scaling via Adversarial Inverse Reinforcement Learning
- Transformers as Bayesian In-Context Experimenters: Smoothness-Adaptive Efficient ATE Estimation
- Convergence issues in Relational Concept Analysis based on AOC-posets
- Beliefs Beyond Posteriors: The Predictive Cost of ELBO Binding
- Understanding Diffusion Models via Ratio-Based Function Approximation with SignReLU Networks