Research map / Label-efficient and robust learning / Knowledge distillation
Predictive Student Modeling in Educational Games with Multi-Task Learning
Michael Geden, Andrew Emerson, Jonathan Rowe, Roger Azevedo, James Lester
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 multi-task learning (one task per post-test question) instead of a single summed-score output (ablation) | stacked LSTM · Crystal Island post-test | MSE: 8.36 -> 6.93 | Improved |
| use sequential LSTM representation instead of static aggregated features | student model · Crystal Island post-test | MSE: 8.69 -> 8.36 (Lasso vs single-task LSTM) | Improved |
| add full self-attention (ablation) | unweighted MTL LSTM · Crystal Island post-test | MSE: 6.93 -> 6.99 | No clear effect |
| add full self-attention (ablation) | uncertainty-weighted MTL LSTM · Crystal Island post-test | MSE: 7.40 -> 6.92 | Improved |
| use uncertainty-weighted loss instead of equal task weights (ablation) | MTL LSTM · Crystal Island post-test | MSE: 6.93 -> 7.40 (no attention) | No clear effect |
| increase number of tasks (ablation) | MTL LSTM · Crystal Island post-test | accuracy/fit | Improved |
Closest papers by meaning
- DataEnvGym: Data Generation Agents in Teacher Environments with Student Feedback (ICLR 2025 · similarity 0.84)
- Can Students Beyond the Teacher? Distilling Knowledge from Teacher’s Bias (AAAI 2025 · similarity 0.84)
- Simulating Students or Sycophantic Problem Solving? On Misconception Faithfulness of LLM Simulators (NeurIPS 2026 · similarity 0.84)
- Teaching Models to Teach Themselves: Reasoning at the Edge of Learnability (ICML 2026 · similarity 0.83)
- Cooperative Knowledge Distillation: A Learner Agnostic Approach (AAAI 2024 · similarity 0.83)
- DistiLLM-2: A Contrastive Approach Boosts the Distillation of LLMs (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.
