singular value decomposition
A matrix factorization technique that decomposes a matrix into its constituent components, aiding in dimensionality reduction, data compression, and revealing underlying patterns in high-dimensional data.
- CURE: Concept Unlearning via Orthogonal Representation Editing in Diffusion Models
- Continuous Subspace Optimization for Continual Learning
- EMLoC: Emulator-based Memory-efficient Fine-tuning with LoRA Correction
- Efficient Parametric SVD of Koopman Operator for Stochastic Dynamical Systems
- FedSVD: Adaptive Orthogonalization for Private Federated Learning with LoRA
- Learning with Restricted Boltzmann Machines: Asymptotics of AMP and GD in High Dimensions
- LoRA vs Full Fine-tuning: An Illusion of Equivalence
- MoORE: SVD-based Model MoE-ization for Conflict- and Oblivion-Resistant Multi-Task Adaptation
- Parameter Efficient Fine-tuning via Explained Variance Adaptation
- SUMO: Subspace-Aware Moment-Orthogonalization for Accelerating Memory-Efficient LLM Training
- StelLA: Subspace Learning in Low-rank Adaptation using Stiefel Manifold