Research map / Label-efficient and robust learning / Recognition and long-tail
Event Representations With Tensor-Based Compositions
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 tensor composition instead of additive or concatenation-based composition | Event representations · Hard Similarity | Accuracy: Role Factor Tensor: 43.5† vs Compositional Neural Net: 33.0; Predicate Tensor: 41.0 † vs Compositional Neural Net: 33.0 | Helped |
| Train event representations to predict words rather than events | Tensor event representations · Transitive Sentence Similarity | Spearman's ρ: Role Factor Tensor: 0.64 -> 0.71; Predicate Tensor: 0.63 -> 0.71 | Helped |
| Train event representations to predict events rather than words | Role Factor Tensor · Hard Similarity | Accuracy: 23.5 -> 43.5 | Helped |
| Train event representations to predict events rather than words | Role Factor Tensor · CMCNC | Accuracy: 70.1 -> 72.1 | Helped |
| Use Role Factor Tensor rather than Compositional Neural Net | Event representations · CMCNC | Accuracy: 65.7 -> 70.1 (WP); 68.5 -> 72.1 (EV) | Helped |
| Use Role Factor Tensor rather than Compositional Neural Net | Event representations · MCNC | Accuracy: 38.4 -> 42.2 (WP); 45.3 -> 46.5 (EV) | Helped |
| Use Predicate Tensor rather than Compositional Neural Net | Event representations · MCNC | Accuracy: 38.4 -> 32.3 (WP); 45.3 -> 41.1 (EV) | Hurt |
| Generate schemas from Role Factor event representations rather than Relgrams counts | Event schema generation · Generated Schemas | Average annotator score: 1.51 -> 2.45 (with 0's); 1.78 -> 2.62 (without 0's) | Helped |
| Use Role Factor event representations rather than Compositional Neural Net representations | Event schema generation · Generated Schemas | Average annotator score: 2.26 -> 2.45 (with 0's); 2.47 -> 2.62 (without 0's) | Helped |
Closest papers by meaning
- FEEL: Featured Event Embedding Learning (AAAI 2018 · similarity 0.87)
- Rich Event Modeling for Script Event Prediction (AAAI 2023 · similarity 0.87)
- A Generative Approach for Script Event Prediction via Contrastive Fine-Tuning (AAAI 2023 · similarity 0.86)
- A Comprehensive Evaluation on Event Reasoning of Large Language Models (AAAI 2025 · similarity 0.84)
- Emergent Representations of Program Semantics in Language Models Trained on Programs (ICML 2024 · similarity 0.82)
- CHRONOS: A Schema-Based Event Understanding and Prediction System (AAAI 2024 · 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.