differentiable optimization
Differentiable optimization refers to optimization methods that involve differentiable functions, allowing gradient-based techniques to be applied. This is prevalent in training AI models, where the loss function's gradients guide the updates to model parameters throughout the learning process.
- Bridging Arbitrary and Tree Metrics via Differentiable Gromov Hyperbolicity
- Differentiable Structure Learning and Causal Discovery for General Binary Data
- Differentiation Through Black-Box Quadratic Programming Solvers
- FSNet: Feasibility-Seeking Neural Network for Constrained Optimization with Guarantees
- Puppeteer: Rig and Animate Your 3D Models
- Template-Guided 3D Molecular Pose Generation via Flow Matching and Differentiable Optimization
- Unsupervised Learning for Optimal Transport plan prediction between unbalanced graphs