gradient-based optimization
A collection of optimization techniques that utilize the gradient of a function to find local minima or maxima. In machine learning, this method is employed to adjust model parameters based on loss functions to improve performance iteratively.
- Bridging Arbitrary and Tree Metrics via Differentiable Gromov Hyperbolicity
- Differentiable Constraint-Based Causal Discovery
- Differentiable Decision Tree via "ReLU+Argmin" Reformulation
- Direct Fisher Score Estimation for Likelihood Maximization
- Fast Projection-Free Approach (without Optimization Oracle) for Optimization over Compact Convex Set
- Generalization Bound of Gradient Flow through Training Trajectory and Data-dependent Kernel
- H3D-DGS: Exploring Heterogeneous 3D Motion Representation for Deformable 3D Gaussian Splatting
- Heavy-Ball Momentum Method in Continuous Time and Discretization Error Analysis
- Information-Driven Design of Imaging Systems
- Precise Diffusion Inversion: Towards Novel Samples and Few-Step Models
- SPARKE: Scalable Prompt-Aware Diversity and Novelty Guidance in Diffusion Models via RKE Score
- Solving the Asymmetric Traveling Salesman Problem via Trace-Guided Cost Augmentation
- Trust Region Constrained Measure Transport in Path Space for Stochastic Optimal Control and Inference
- VERA: Variational Inference Framework for Jailbreaking Large Language Models