Research map / Label-efficient and robust learning / Knowledge distillation
Proposition Entailment in Educational Applications using Deep Neural Networks
Florin Bulgarov, Rodney Nielsen
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 50-dim word embeddings with DNN instead of manual features with SVM (ablation) | MMP entailment · MMP entailment test | weighted F1: 0.73 -> 0.76 | Improved |
| use word embeddings instead of manual features in SVM (ablation) | SVM · MMP entailment test | F1 understood class: 0.53 -> 0.6 | Improved |
| add manual features to SVM with WEs (ablation) | SVM (WEs) · MMP entailment test | weighted F1: 0.75 -> 0.76 | No clear effect |
| add manual features to DNN with WEs (ablation) | DNN (WEs) · MMP entailment test | F1 understood class: 0.63 -> 0.55 | Hurt |
| add LSA score to SVM with manual features (ablation) | SVM (man. ftrs.) · MMP entailment test | weighted F1: 0.73 -> 0.73 | No clear effect |
| add LSA score to DNN with WEs (ablation) | DNN (WEs) · MMP entailment test | weighted F1: 0.76 -> 0.66 | Hurt |
| add Corley & Mihalcea score to DNN with WEs (ablation) | DNN (WEs) · MMP entailment test | weighted F1: 0.76 -> 0.65 | Hurt |
| add both LSA and C&M scores to DNN with WEs (ablation) | DNN (WEs) · MMP entailment test | weighted F1: 0.76 -> 0.66 | Hurt |
| increase word embedding dimension from 50 to 100 (ablation) | DNN (WEs) · MMP entailment test | weighted F1: 0.76 -> 0.73 | Hurt |
| increase word embedding dimension from 50 to 200 (ablation) | DNN (WEs) · MMP entailment test | weighted F1: 0.76 -> 0.74 | Hurt |
| use DNN with manual features instead of SVM (ablation) | manual features · MMP entailment test | weighted F1: 0.73 -> 0.72 | No clear effect |
| replace Horbach et al. state-of-the-art system with WE DNN (ablation) | MMP entailment · MMP entailment test | F1 understood class: 0.39 -> 0.63 | Improved |
Closest papers by meaning
- Can Neural Networks Understand Logical Entailment? (ICLR 2018 · similarity 0.87)
- SciTaiL: A Textual Entailment Dataset from Science Question Answering (AAAI 2018 · similarity 0.85)
- Enhancing Student Performance Prediction on Learnersourced Questions with SGNN-LLM Synergy (AAAI 2024 · similarity 0.84)
- A Context-Enriched Neural Network Method for Recognizing Lexical Entailment (AAAI 2017 · similarity 0.84)
- Automated Assessment of Student Self-Explanation in Code Comprehension Using Pre-Trained Language Models (AAAI 2025 · similarity 0.84)
- EduLLM: Leveraging Large Language Models and Framelet-Based Signed Hypergraph Neural Networks for Student Performance Prediction (ICML 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.
