Research map / Label-efficient and robust learning / Semi-supervised learning
DL2: Training and Querying Neural Networks with Logic
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 |
|---|---|---|---|
| add DL2 group-probability constraint on unlabeled CIFAR-100 data (semi-supervised) | CIFAR-100 classifier · Semi-supervised CIFAR-100 | Accuracy (%): +2.2 (45.20 to 47.40) | Helped |
| add DL2 group-probability constraint on unlabeled data (semi-supervised) | CIFAR-100 classifier · Semi-supervised CIFAR-100 | Constraint accuracy (%): 74.80 to 90.85 | Helped |
| add global Lipschitz constraint via DL2 query-based adversary | MNIST classifier · LipschitzG | Constraint accuracy (%): 0.00 to 100.00 | Helped |
| add global Lipschitz constraint | FASHION classifier · LipschitzG | Prediction accuracy (%): 92.79 to 81.10 | Hurt |
| add global robustness constraint | MNIST, FASHION, CIFAR-10 classifiers · RobustnessG | Constraint accuracy (%): 0.00 to 99.96 (MNIST), 0.00 to 99.91 (FASHION), 0.00 to 99.95 (CIFAR-10) | Helped |
| add training-set Lipschitz constraint | MNIST classifier · LipschitzT | Constraint accuracy (%): 9.65 to 99.90 | Helped |
| add training-set Lipschitz constraint | MNIST classifier · LipschitzT | Prediction accuracy (%): 99.54 to 97.90 | Hurt |
| add global segment constraint | MNIST classifier · SegmentG | Constraint accuracy (%): 17.58 to 42.19 | Helped |
| add global C-similarity constraint | CIFAR-10 classifier · C-similarityG | Constraint accuracy (%): 50.47 to 99.42 | Helped |
| add global C-similarity constraint | CIFAR-10 classifier · C-similarityG | Prediction accuracy (%): 91.89 to 88.99 | Hurt |
| add training-set constraints (unsupervised graph distance via DL2 loss, no labels) | MLP on graph min-distance · Unsupervised graph min-distance regression | MSE: 0.6680 to 0.1030 | Helped |
Closest papers by meaning
- LogicMP: A Neuro-symbolic Approach for Encoding First-order Logic Constraints (ICLR 2024 · similarity 0.86)
- Mind the Gap: Removing the Discretization Gap in Differentiable Logic Gate Networks (NeurIPS 2025 · similarity 0.84)
- An Evaluation of Approaches to Train Embeddings for Logical Inference (Student Abstract) (AAAI 2025 · similarity 0.84)
- DeepSaDe: Learning Neural Networks That Guarantee Domain Constraint Satisfaction (AAAI 2024 · similarity 0.84)
- Neural Reasoning Networks: Efficient Interpretable Neural Networks with Automatic Textual Explanations (AAAI 2025 · similarity 0.84)
- On the Hardness of Training Deep Neural Networks Discretely (AAAI 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.