Differentiable Knapsack and Top-k Operators via Dynamic Programming
cs.LG
Submitted: 2026-01-29
Updated: 2026-11-05
Terminology
Sources
- SIMPLE: A Gradient Estimator for $k$-Subset Sampling
- Estimating or Propagating Gradients Through Stochastic Neurons for Conditional Computation
- Learning with Differentiable Perturbed Optimizers
- The Elements of Differentiable Programming
- Learning Classifiers with Fenchel-Young Losses: Generalized Entropies, Margins, and Algorithms
- Learning with Fenchel-Young Losses
- Fast Differentiable Sorting and Ranking
- Differentiable Patch Selection for Image Recognition
- Differentiable Ranks and Sorting using Optimal Transport
- MIPaaL: Mixed Integer Program as a Layer
- Structured Reinforcement Learning for Combinatorial Decision-Making
- Stochastic Beams and Where to Find Them: The Gumbel-Top-k Trick for Sampling Sequences Without Replacement
- Interior Point Solving for LP-based prediction+optimisation
- From Softmax to Sparsemax: A Sparse Model of Attention and Multi-Label Classification
- Differentiable Dynamic Programming for Structured Prediction and Attention
- Contrastive Losses and Solution Caching for Predict-and-Optimize
- Implicit MLE: Backpropagating Through Discrete Exponential Family Distributions
- Differentiable Sorting Networks for Scalable Sorting and Ranking Supervision
- Multi-Vector Retrieval as Sparse Alignment
- Discrete Variational Autoencoders
Related papers
- A Physics-Regulated Neural Framework for Learning 3D Grain Growth Dynamics
- FlexMoE: Learning One-for-All Multi-Budget Expert Pruning for Flexible MoE Deployment
- OptiMind: Improving Small Language Models for Optimization Modeling through Formulation Guidance and Multi-Turn Learning
- Staying in the Sweet Spot: Boost Reasoning Evolution via Responsive Rollout Difficulty Manipulation
- Examining the Impact of Data and Training Methods on Emergent and Subliminal Misalignment
- Rethinking Black-Box On-Policy Distillation: A Rubric-based Analysis