approximation quality
Approximation quality refers to how well a simplified or approximated model represents the true complexities of a given problem, often assessed in the context of approaches that aim to reduce computational cost while maintaining adequate performance.
- Breaking the Frozen Subspace: Importance Sampling for Low-Rank Optimization in LLM Pretraining
- COALA: Numerically Stable and Efficient Framework for Context-Aware Low-Rank Approximation
- Degrees of Freedom for Linear Attention: Distilling Softmax Attention with Optimal Feature Efficiency
- Final-Model-Only Data Attribution with a Unifying View of Gradient-Based Methods
- Private Geometric Median in Nearly-Linear Time
- Streaming Stochastic Submodular Maximization with On-Demand User Requests