Research map / Architectures and efficiency / Neural architecture search
Efficient Verification of ReLU-Based Neural Networks via Dependency Analysis
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 |
|---|---|---|---|
| Venus (dependency analysis + splitting + ideal cuts) vs NSVerify plain MILP | MILP verification · MNIST | total time: 26,906.81 -> 5,953.46 | Helped |
| add dependency cuts to big-M MILP (ablation) | Venus · MNIST | total time: 13,555.49 -> 9,729.07 | Helped |
| add ideal cuts to big-M MILP (ablation) | Venus · MNIST | total time: 13,555.49 -> 6,777.95 | Helped |
| add input domain splitting on high-dimensional input (ablation) | Venus · MNIST | total time: 13,555.49 -> 11,409.96 | No clear effect |
| combine dependency cuts with ideal cuts (ablation) | Venus · MNIST | total time: 6,777.95 -> 5,953.46 | Helped |
| dependency analysis callback frequency too high or too low (ablation) | Venus · MNIST | average time | Hurt |
Closest papers by meaning
- Dynamic Back-Substitution in Bound-Propagation-Based Neural Network Verification (AAAI 2025 · similarity 0.88)
- Clip-and-Verify: Linear Constraint-Driven Domain Clipping for Accelerating Neural Network Verification (NeurIPS 2025 · similarity 0.86)
- SEEV: Synthesis with Efficient Exact Verification for ReLU Neural Barrier Functions (NeurIPS 2024 · similarity 0.85)
- Provably Safe Neural Network Controllers via Differential Dynamic Logic (NeurIPS 2024 · similarity 0.84)
- Training Verification-Friendly Neural Networks via Neuron Behavior Consistency (AAAI 2025 · similarity 0.84)
- Verifying Properties of Binary Neural Networks Using Sparse Polynomial Optimization (ICLR 2025 · similarity 0.83)
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.