Research map / Vision-language and multimodal / Visual question answering
Commonsense in Parts: Mining Part-Whole Relations from the Web and Image Tags
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 weighted-support and discriminative-strength scoring for assertion ranking | PWKB pattern extraction · assertion ranking | precision: 0.55 -> 0.85 | Helped |
| use IMS-based head-word disambiguation instead of most-frequent-sense | PWKB · argument disambiguation | precision: 0.70 -> 0.80 | Helped |
| rank noun phrases by NPMI plus frequency instead of NPMI alone | WordNet extension · noun phrase quality | precision: 0.25 -> 0.60 | Helped |
| add German and Italian corpora for cardinality inference (ablation) | PWKB cardinality · cardinality | precision: 0.61 -> 0.80 | Helped |
| add transitivity and inheritance deduction rules (ablation) | PWKB · KB size | assertions: 382K -> 6.75M | Helped |
| add irreflexivity and acyclicity constraint checking (ablation) | PWKB · KB consistency | false assertions removed: -146K | Helped |
| use Flickr tags instead of object detector for visibility co-occurrence | visibility attribute · 100K Flickr images | positive matches: ca. 12,000 -> ca. 26,000 | Helped |
| replace Yahoo-snippet part-whole relations with PWKB | zero-shot image classification · 200 unseen categories | top-5 accuracy: 23.8% -> 25.5% | Helped |
Closest papers by meaning
- Joint Commonsense and Relation Reasoning for Image and Video Captioning (AAAI 2020 · similarity 0.88)
- Understanding Multi-Granularity for Open-Vocabulary Part Segmentation (NeurIPS 2024 · similarity 0.84)
- Half-Truths Break Similarity-Based Retrieval (NeurIPS 2026 · similarity 0.84)
- CORECODE: A Common Sense Annotated Dialogue Dataset with Benchmark Tasks for Chinese Large Language Models (AAAI 2024 · similarity 0.84)
- VisMin: Visual Minimal-Change Understanding (NeurIPS 2024 · similarity 0.83)
- PARTONOMY: Large Multimodal Models with Part-Level Visual Understanding (NeurIPS 2025 · similarity 0.83)
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.