Research map / Large language models / Retrieval and knowledge
Faithful Relational Reasoning with Region-based Embeddings: Expressivity of Convex Coordinate-wise Models
Victor Charpenay, Steven Schockaert
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 ExpressivE instead of Octagons | region-based embeddings · FAIRE PERM3 | hits@1: 0.58 -> 0.97 | Improved |
| use ExpressivE instead of Octagons | region-based embeddings · FAIRE MIX3 | hits@1: 0.25 -> 0.71 | Improved |
| use ExpressivE instead of Octagons | region-based embeddings · FAIRE MIX2 | hits@1: 0.47 -> 0.94 | Improved |
| use polygon regions (6 edges) instead of ExpressivE | region-based embeddings · FAIRE PERM3 | hits@1: 0.97 -> 0.69 | Hurt |
| use polygon regions (6 edges) instead of ExpressivE | region-based embeddings · FAIRE COMB | hits@1: 0.32 -> 0.40 | Improved |
| increase polygon edges from 2 to 10 (ablation) | polygon region model · FAIRE PERM3 | hits@1: 0.60 -> 0.89 | Improved |
| increase polygon edges from 2 to 10 (ablation) | polygon region model · FAIRE MIX3 | hits@1: 0.50 -> 0.47 | No clear effect |
| increase embedding dimensionality (ablation) | region-based models · FAIRE | hits@1 | No clear effect |
Closest papers by meaning
- When can transformers reason with abstract symbols? (ICLR 2024 · similarity 0.87)
- Systematic Relational Reasoning With Epistemic Graph Neural Networks (ICLR 2025 · similarity 0.87)
- Improving Soft Unification with Knowledge Graph Embedding Methods (ICML 2025 · similarity 0.87)
- An Evaluation of Approaches to Train Embeddings for Logical Inference (Student Abstract) (AAAI 2025 · similarity 0.87)
- Geometric Latent Reasoning Induces Shorter Generations in LLMs (NeurIPS 2026 · similarity 0.86)
- When No Paths Lead to Rome: Benchmarking Systematic Neural Relational Reasoning (NeurIPS 2025 · similarity 0.86)
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.
