Research map / Detection and segmentation / CNN features
Complementing Semantic Roles with Temporally Anchored Spatial Knowledge: Crowdsourced Annotations and Experiments
Alakananda Vempala, Eduardo Blanco
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 |
|---|---|---|---|
| use machine-learned features instead of most-likely-label baseline | spatial knowledge classifier · All instances | overall F-measure: 0.31 -> 0.52 | Improved |
| add lexical features over baseline (ablation) | classifier · All instances | overall F-measure: 0.31 -> 0.48 | Improved |
| add syntactic features to lexical features (ablation) | classifier · All instances | overall F-measure: 0.48 -> 0.47 | No clear effect |
| add semantic features to lexical + syntactic (ablation) | classifier · All instances | overall F-measure: 0.48 -> 0.52 | Improved |
| add semantic and syntactic features on top of lexical (ablation) | classifier · Day before instances | F-measure: 0.38 -> 0.41 | Improved |
| add semantic and syntactic features on top of lexical (ablation) | classifier · Day after instances | F-measure: 0.41 -> 0.48 | Improved |
| vary feature set (ablation) | classifier · During instances | F-measure: 0.61 -> 0.61 | No clear effect |
| restrict evaluation to instances with a majority label | classifier · All features | overall F-measure: 0.52 -> 0.63 | Improved |
| generate and filter instances with predicted roles (predicted ∩ gold) instead of gold (ablation) | classifier · All features | overall F-measure: 0.52 -> 0.55 | Improved |
| use fully predicted annotations instead of gold (ablation) | classifier · All features | overall F-measure: 0.52 -> 0.29 | Hurt |
| use fully predicted annotations instead of gold (ablation) | classifier · All features | recall: 0.54 -> 0.20 | Hurt |
| use fully predicted annotations instead of gold (ablation) | classifier · All features | precision: 0.51 -> 0.58 | Mixed |
Closest papers by meaning
- EgoMind: Activating Spatial Cognition through Linguistic Reasoning in MLLMs (CVPR 2026 · similarity 0.83)
- SpatialScore: Towards Comprehensive Evaluation for Spatial Intelligence (CVPR 2026 · similarity 0.82)
- Is A Picture Worth A Thousand Words? Delving Into Spatial Reasoning for Vision Language Models (NeurIPS 2024 · similarity 0.82)
- Stitch and Tell: A Structured Data Augmentation Method for Spatial Understanding (NeurIPS 2025 · similarity 0.82)
- Linear Mechanisms for Spatiotemporal Reasoning in Vision Language Models (ICLR 2026 · similarity 0.82)
- Unaligned Message-Passing and Contextualized-Pretraining for Robust Geo-Entity Resolution (AAAI 2025 · similarity 0.82)
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.
