Research map / Graph learning / Knowledge and temporal graphs
Graph-Based Reasoning over Heterogeneous External Knowledge for Commonsense Question Answering
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 topology sort to reorder evidence using graph structure (ablation) | XLNet + E · CommonsenseQA dev | accuracy: 75.8 -> 77.7; 1.9% | Helped |
| Add graph inference with GCN and graph attention (ablation) | XLNet + E · CommonsenseQA dev | accuracy: 75.8 -> 77.2; 1.4% | Helped |
| Add topology sort and graph inference together (ablation) | XLNet + E · CommonsenseQA dev | accuracy: 75.8 -> 79.3; 3.5% | Helped |
| Add ConceptNet evidence and graph-based reasoning (ablation) | XLNet-large · CommonsenseQA dev | accuracy: 68.9 -> 75.3; 6.4% | Helped |
| Add Wikipedia evidence and graph-based reasoning (ablation) | XLNet-large · CommonsenseQA dev | accuracy: 68.9 -> 73.5; 4.6% | Helped |
| Add both ConceptNet and Wikipedia evidence with graph-based reasoning (ablation) | XLNet-large · CommonsenseQA dev | accuracy: 68.9 -> 79.3; 9.4% | Helped |
| Combine ConceptNet and Wikipedia rather than using either source alone (ablation) | XLNet-large with graph-based reasoning · CommonsenseQA dev | accuracy: ConceptNet: 75.3; Wikipedia: 73.5; ConceptNet + Wikipedia: 79.3 | Helped |
| Use heterogeneous evidence and graph-based reasoning instead of XLNet-large alone | Our Model · CommonsenseQA blind test | accuracy: 62.9 -> 75.3; 12.4% | Helped |
| Use heterogeneous evidence and graph-based reasoning instead of RoBERTa(ensemble) | Our Model · CommonsenseQA blind test | accuracy: 72.5 -> 75.3; 2.8% | Helped |
Closest papers by meaning
- Dynamic Neuro-Symbolic Knowledge Graph Construction for Zero-shot Commonsense Question Answering (AAAI 2021 · similarity 0.89)
- Debate on Graph: A Flexible and Reliable Reasoning Framework for Large Language Models (AAAI 2025 · similarity 0.88)
- Amplifying Diversity and Quality in Commonsense Knowledge Graph Completion (Student Abstract) (AAAI 2024 · similarity 0.88)
- Multimodal Commonsense Knowledge Distillation for Visual Question Answering (Student Abstract) (AAAI 2025 · similarity 0.88)
- Graph Reasoning Transformers for Knowledge-Aware Question Answering (AAAI 2024 · similarity 0.86)
- KG-Guard: Graph-Based Hallucination Detection for Knowledge Base Question Answering (NeurIPS 2026 · 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.