Research map / Label-efficient and robust learning / Recognition and long-tail
Combining Fact Extraction and Verification with Neural Semantic Matching Networks
Yixin Nie, Haonan Chen, Mohit Bansal
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 |
|---|---|---|---|
| Add dNSMN reranking after keyword matching | FEVER document retrieval · FEVER dev, k=5 | OFEVER: 88.86 -> 92.34 | Improved |
| Add Pageview ranking to KM + dNSMN | FEVER document retrieval · FEVER dev, k=5 | OFEVER: 92.34 -> 92.42 | Improved |
| Add Pageview ranking to KM + dNSMN | FEVER document retrieval · FEVER dev, k=10 | OFEVER: 92.82 -> 92.75 | No clear effect |
| Add annealed sampling during training (ablation) | sNSMN sentence selection · FEVER dev | OFEVER: 86.65 -> 91.19 | Improved |
| Add annealed sampling during training (ablation) | sNSMN sentence selection · FEVER dev | F1: 74.33 -> 51.38 | Hurt |
| Add WordNet and number features (ablation) | vNSMN verification · FEVER dev | FEVER: 65.37 -> 66.14 | Improved |
| Add sentence-selection semantic relatedness scores (ablation) | vNSMN verification · FEVER dev | FEVER: 64.90 -> 66.14 | Improved |
| Add document retrieval relatedness scores as well as sentence-selection scores (ablation) | vNSMN verification · FEVER dev | FEVER: 66.14 -> 66.05 | Hurt |
| Replace the proposed verifier with vanilla ESIM (ablation) | FEVER verification · FEVER dev | FEVER: 66.14 -> 65.07 | Hurt |
| Use TF-IDF-selected data for non-verifiable examples instead of sNSMN-selected data (ablation) | vNSMN verification · FEVER dev | FEVER: 66.14 -> 62.48 | Hurt |
| Lower the evidence filtering threshold from 0.5 to 0.05 (ablation) | vNSMN verification · FEVER dev | FEVER: 66.15 -> 66.49 | Improved |
| Add evidence enhancement using linked documents (ablation) | full FEVER system · FEVER dev | FEVER: +0.08% | Improved |
Closest papers by meaning
- MRR-FV: Unlocking Complex Fact Verification with Multi-Hop Retrieval and Reasoning (AAAI 2025 · similarity 0.87)
- Heterogeneous Graph Reasoning for Fact Checking over Texts and Tables (AAAI 2024 · similarity 0.86)
- Enhancing Multi-Hop Fact Verification with Structured Knowledge-Augmented Large Language Models (AAAI 2025 · similarity 0.85)
- Unsupervised Pretraining for Fact Verification by Language Model Distillation (ICLR 2024 · similarity 0.85)
- AVerImaTeC: A Dataset for Automatic Verification of Image-Text Claims with Evidence from the Web (NeurIPS 2025 · similarity 0.85)
- Beyond Text: Fine-Grained Multi-Modal Fact Verification with Hypergraph Transformers (AAAI 2025 · similarity 0.84)
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.
