Research map / Learning theory and algorithms / Combinatorial and search optimization
Deep FlexQP: Accelerated Nonlinear Programming via Deep Unfolding
Alex Oshin, Rahul Vodeb Ghosh, Augustinos D Saravanos, Evangelos Theodorou
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| use Deep FlexQP inside SQP instead of OSQP | SQP · nonlinear trajectory optimization | solve time: 4-16x faster | Improved |
| use Deep FlexQP inside SQP instead of OSQP | SQP · highly constrained NLPs (100 problems) | solve time: over 16x faster | Improved |
| use Deep FlexQP instead of existing methods | predictive safety filter · predictive safety filter | safety violations: over 70% | Improved |
| use Deep FlexQP instead of existing methods | predictive safety filter · predictive safety filter | task completion: 43% | Improved |
| fine-tune Deep OSQP variants with Lagrange multiplier loss instead of optimality gap loss (ablation) | Deep OSQP · large-scale SVM QPs (10k variables) | convergence | Hurt |
| fine-tune with Lagrange multiplier loss vs optimality gap loss (ablation) | Deep OSQP · large-scale portfolio QPs | iterations | No clear effect |
| use log-scaled PAC-Bayes loss (ablation) | Deep FlexQP · LASSO (1000 test problems) | generalization bound informativeness | Improved |
| use fine-tuning on large-scale problems instead of training from scratch | Deep FlexQP · large-scale QPs | training time: over 300 days | Improved |
| use LSTM policies (ablation) | Deep FlexQP · problems with changing active constraints | convergence | Improved |
Closest papers by meaning
- Deep Distributed Optimization for Large-Scale Quadratic Programming (ICLR 2025 · similarity 0.90)
- BayeSQP: Bayesian Optimization through Sequential Quadratic Programming (NeurIPS 2025 · similarity 0.87)
- Differentiation Through Black-Box Quadratic Programming Solvers (NeurIPS 2025 · similarity 0.85)
- OptiBench Meets ReSocratic: Measure and Improve LLMs for Optimization Modeling (ICLR 2025 · similarity 0.85)
- BPQP: A Differentiable Convex Optimization Framework for Efficient End-to-End Learning (NeurIPS 2024 · similarity 0.85)
- Principled Data Augmentation for Learning to Solve Quadratic Programming Problems (NeurIPS 2025 · similarity 0.85)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
