Research map / Learning theory and algorithms / Combinatorial and search optimization
Learning Interpretable Temporal Properties from Positive Examples Only
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 |
|---|---|---|---|
| Replace counterexample-guided search with symbolic DFA language-minimality constraints | DFA learning · 28 random DFAs, size 2 to 10 | main-loop iterations: average ratio of the number of iterations is $0.14$ | Helped |
| Replace counterexample-guided search with symbolic DFA language-minimality constraints | DFA learning · 28 random DFAs, size 2 to 10 | inference time: average ratio of the inference time is $1.09$ | No clear effect |
| Combine symbolic constraints with generated negative examples | LTL_f learning · 12 common LTL_f pattern samples | inference time: 173.9% faster | Helped |
| Use semi-symbolic learning rather than counterexample-guided learning | LTL_f learning · 12 common LTL_f pattern samples | samples with inferred formulas equally or more specific than ground truth: three out of the 12 -> five out of the 12 | Helped |
| Combine symbolic constraints with generated negative examples | LTL_f learning · UAV trajectory clusters | inference time: 260.73% faster | Helped |
| Use a semi-symbolic approach for DFA learning instead of the symbolic approach (ablation) | DFA learning · DFA learning comparison | inference time: 3.2 times more | Hurt |
| Use a semi-symbolic approach for DFA learning instead of the symbolic approach (ablation) | DFA learning · DFA learning comparison | main-loop iterations: 2.1 times more | Hurt |
| Use completely symbolic LTL_f learning instead of semi-symbolic learning (ablation) | LTL_f learning · LTL_f pattern samples | inference time: several orders of magnitude above | Hurt |
| Use completely symbolic LTL_f learning (ablation) | LTL_f learning · UAV trajectory clusters | completion within timeout: timed out even for $n=1$ | Hurt |
Closest papers by meaning
- First-Order Automata (AAAI 2025 · similarity 0.84)
- End-to-End Learning of LTLf Formulae by Faithful LTLf Encoding (AAAI 2024 · similarity 0.84)
- STL: Still Tricky Logic (for System Validation, Even When Showing Your Work) (NeurIPS 2024 · similarity 0.84)
- Neuro-Symbolic Temporal Point Processes (ICML 2024 · similarity 0.83)
- Compositional Automata Embeddings for Goal-Conditioned Reinforcement Learning (NeurIPS 2024 · similarity 0.83)
- DeepLTL: Learning to Efficiently Satisfy Complex LTL Specifications for Multi-Task RL (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.