Research map / Large language models / AI and society
Aligning Artificial Neural Networks and Ontologies towards Explainable AI
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 |
|---|---|---|---|
| Extract ontology concepts relevant to the main network's classification task from its activations | NNA, NNB, and NNC · Train-image concept extraction | accuracy | Helped |
| Use main-network activations to train mapping networks instead of training convolutional networks on images | Concept extraction · Relevant train concepts | accuracy | Helped |
| Use Input Reduce-selected features instead of all dense-part activations (ablation) | Mapping networks for train concepts · Table 1 concept extraction | accuracy: 0.9367 -> 0.9263; 0.9719 -> 0.9930; 0.9937 -> 0.9942; 0.9950 -> 0.9928; 0.9676 -> 0.9629; 0.9485 -> 0.9433; 0.9670 -> 0. | Mixed |
| Use Input Reduce-selected features instead of all dense-part activations (ablation) | Mapping networks for train concepts · Table 1 feature count | number of features: 10480 -> 453; 10480 -> 4; 10480 -> 2; 10480 -> 4; 10464 -> 2374; 10464 -> 1107; 10464 -> 534; 10464 -> 1247; 10608 -> 51 | Helped |
| Map extracted concepts through ontology axioms to generate symbolic justifications | NNA, NNB, and NNC · Train images matching each network's target type | all justifications correct: 85.5%; 94.2%; 90.6% | Helped |
Closest papers by meaning
- Concept frustration: Aligning human concepts and machine representations (NeurIPS 2026 · similarity 0.84)
- Neurons to Words: A Novel Method for Automated Neural Network Interpretability and Alignment (AAAI 2025 · similarity 0.84)
- Neuro-Symbolic Integration for Reasoning and Learning on Knowledge Graphs (AAAI 2024 · similarity 0.84)
- Learning Visually Grounded Domain Ontologies via Embodied Conversation and Explanation (AAAI 2025 · similarity 0.83)
- Keep the Faith: Faithful Explanations in Convolutional Neural Networks for Case-Based Reasoning (AAAI 2024 · similarity 0.82)
- FaCT: Faithful Concept Traces for Explaining Neural Network Decisions (NeurIPS 2025 · similarity 0.82)
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.