constraint satisfaction
Constraint satisfaction in AI involves solving problems that require finding values for variables subject to constraints or restrictions, which is a common framework in optimization and combinatorial methods.
- BikeBench: A Bicycle Design Benchmark for Generative Models with Objectives and Constraints
- Constrained Diffusers for Safe Planning and Control
- Don’t Trade Off Safety: Diffusion Regularization for Constrained Offline RL
- Efficient Safe Meta-Reinforcement Learning: Provable Near-Optimality and Anytime Safety
- FSNet: Feasibility-Seeking Neural Network for Constrained Optimization with Guarantees
- Gradient-Guided Epsilon Constraint Method for Online Continual Learning
- InstructFlow: Adaptive Symbolic Constraint-Guided Code Generation for Long-Horizon Planning
- MESS+: Dynamically Learned Inference-Time LLM Routing in Model Zoos with Service Level Guarantees