Research map / Bandits and online learning / Multi-armed bandits
Machine Teaching of Active Sequential Learners
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 a planning teacher instead of a naive teacher | Teaching-unaware Bayesian Bernoulli bandit learner · Word dataset (main), simulated bandits | expected cumulative reward | Helped |
| Use a planning teacher instead of a naive teacher | Teaching-unaware Bayesian Bernoulli bandit learner · Word dataset (main), simulated bandits | concordance index | Helped |
| Model the planning teacher when learning from its responses | Bayesian Bernoulli bandit with a planning teacher · Word dataset (main), simulated bandits | expected cumulative reward | Helped |
| Model the planning teacher when learning from its responses | Bayesian Bernoulli bandit with a planning teacher · Word dataset (main), simulated bandits | concordance index | Helped |
| Expect a planning teacher when the actual teacher is naive | Planning-aware Bayesian Bernoulli bandit learner · Word dataset (main), simulated bandits | performance | Hurt |
| Use a mixture of naive and planning teacher models | Bayesian Bernoulli bandit learner · Word dataset (main), simulated bandits | performance | Helped |
| Use a mixture teacher model instead of a naive teacher model | Bandit learner interacting with human participants · User study, 20 target words | average cumulative reward: p-value < 0.01 | Helped |
| Use a planning teacher for active-learning queries | Uncertainty-sampling logistic regression learner · Wine Quality test set | classification accuracy | Helped |
| Increase planning horizon from 1 step to 2–4 steps (ablation) | Matched planning teacher–learner pair (P-P) · Word dataset, simulated bandits | cumulative reward | Helped |
| Reduce the planning teacher's optimality parameter (ablation) | Simulated planning teacher · Word dataset, simulated bandits | performance | Hurt |
| Make the reward profile more informative, with more high-reward arms (ablation) | Planning teacher steering a naive learner · Word dataset, supplementary reward profile | performance gain from planning | Hurt |
| Increase the number of arms from 100 to 500 (ablation) | Simulated bandit teaching · Word dataset, supplementary setting | performance | No clear effect |
Closest papers by meaning
- Teaching to Learn: Sequential Teaching of Learners with Internal States (AAAI 2023 · similarity 0.89)
- Teaching with Limited Information on the Learner’s Behaviour (ICML 2020 · similarity 0.89)
- Iterative Classroom Teaching (AAAI 2019 · similarity 0.88)
- Gradient-Based Algorithms for Machine Teaching (CVPR 2021 · similarity 0.88)
- Preference-Based Batch and Sequential Teaching: Towards a Unified View of Models (NeurIPS 2019 · similarity 0.88)
- On the Complexity of Teaching a Family of Linear Behavior Cloning Learners (NeurIPS 2024 · similarity 0.85)
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.