Research map / Label-efficient and robust learning / Recognition and long-tail
Neuro-Symbolic Hierarchical Rule Induction
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 |
|---|---|---|---|
| Use generic hierarchical rule induction with the same proto-rule set across ILP tasks | HRI · Classic ILP tasks | successful runs (%): 100; 40; 0 | Mixed |
| Use HRI rather than the strongest reported baseline | Adjacent to Red · Adjacent to Red | successful runs (%): 100 -> 100 | No clear effect |
| Use HRI rather than the strongest reported baseline | Two Children · Two Children | successful runs (%): 95 -> 100 | Helped |
| Use HRI rather than the strongest reported baseline | Even-Succ2 · Even-Succ2 | successful runs (%): 100 -> 40 | Hurt |
| Use HRI rather than the strongest reported baseline | Buzz · Buzz | successful runs (%): 70 -> 40 | Hurt |
| Use HRI rather than the strongest reported baseline | Visual Genome (GQA), 150-object classification · Visual Genome (GQA), 150-object classification | R@1: 0.53 -> 0.53 | No clear effect |
| Use HRI rather than NLIL | Visual Genome (GQA), 150-object classification · Visual Genome (GQA), 150-object classification | R@5: 0.52 -> 0.60 | Helped |
| Initialize predicate embeddings with NLIL pretrained embeddings instead of random embeddings (ablation) | HRI · Visual Genome, Car task | Accuracy (soft evaluation): 0.63 -> 0.75 | Helped |
| Initialize predicate embeddings with GPT2 pretrained embeddings instead of random embeddings (ablation) | HRI · Visual Genome, Car task | Accuracy (soft evaluation): 0.63 -> 0.65 | Helped |
| Initialize predicate embeddings with NLIL pretrained embeddings instead of random embeddings (ablation) | HRI · Visual Genome, Car task | R@1 (symbolic evaluation): 0.38 -> 0.58 | Helped |
| Initialize predicate embeddings with GPT2 pretrained embeddings instead of random embeddings (ablation) | HRI · Visual Genome, Car task | R@1 (symbolic evaluation): 0.38 -> 0.5 | Helped |
Closest papers by meaning
- Neuro-Symbolic Temporal Point Processes (ICML 2024 · similarity 0.89)
- Lifting Traces to Logic: Programmatic Skill Induction with Neuro-Symbolic Learning for Long-Horizon Agentic Tasks (ICML 2026 · similarity 0.87)
- Right for the Right Reasons: Avoiding Reasoning Shortcuts via Prototypical Neurosymbolic AI (NeurIPS 2025 · similarity 0.86)
- Constraints-Guided Diffusion Reasoner for Neuro-Symbolic Learning (AAAI 2026 · similarity 0.86)
- Position: The Road to Generalizable Neuro-Symbolic Learning Should be Paved with Foundation Models (NeurIPS 2026 · similarity 0.86)
- Mastering Symbolic Operations: Augmenting Language Models with Compiled Neural Networks (ICLR 2024 · 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.