Research map / Label-efficient and robust learning / Recognition and long-tail
Better Set Representations For Relational Reasoning
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 SRN before Relational Network | RN · Sort-of-CLEVR Furthest | accuracy (%): 91.4 -> 95.9 | Helped |
| add SRN before Relational Network | RN · Sort-of-CLEVR Closest | accuracy (%): 93.9 -> 94.8 | Helped |
| add SRN before Relational Network | RN · Sort-of-CLEVR Count | accuracy (%): 100 -> 100 | No clear effect |
| add SRN before Relational Network | RN · Sort-of-CLEVR relational questions | accuracy (%): 95.1 -> 96.9 | Helped |
| add SRN before Relational Network | RN · Sort-of-CLEVR synthetic robustness | robustness (%): 1.5 -> 96.7 | Helped |
| add SRN before Relational Network | RN · Sort-of-CLEVR dataset robustness | robustness (%): 61.3 -> 93.2 | Helped |
| evaluate SRN trained with 5 inner steps using 0 inner steps (ablation) | SRN · Sort-of-CLEVR Furthest | accuracy (%): 96 -> 63 | Hurt |
| evaluate SRN trained with 5 inner steps using 0 inner steps (ablation) | SRN · Sort-of-CLEVR Closest | accuracy (%): 95 -> 87 | Hurt |
| evaluate SRN trained with 5 inner steps using 0 inner steps (ablation) | SRN · Sort-of-CLEVR Count | accuracy (%): 100 -> 100 | No clear effect |
| add SRN to C-SWM with 5 slots | C-SWM · Space Invaders, 1 step | H@1 (%): 54.8 -> 64.7 | Helped |
| add SRN to C-SWM with 5 slots | C-SWM · Space Invaders, 10 steps | H@1 (%): 27.1 -> 36.6 | Helped |
| add SRN to C-SWM with 3 slots | C-SWM · Space Invaders, 1 step | H@1 (%): 69.4 -> 69.0 | No clear effect |
Closest papers by meaning
- An Explicitly Relational Neural Network Architecture (ICML 2020 · similarity 0.89)
- When No Paths Lead to Rome: Benchmarking Systematic Neural Relational Reasoning (NeurIPS 2025 · similarity 0.88)
- Systematic Relational Reasoning With Epistemic Graph Neural Networks (ICLR 2025 · similarity 0.88)
- Neural Reasoning for Sure Through Constructing Explainable Models (AAAI 2025 · similarity 0.86)
- When can transformers reason with abstract symbols? (ICLR 2024 · similarity 0.86)
- Relational Concept Bottleneck Models (NeurIPS 2024 · similarity 0.85)
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.