Research map / Learning theory and algorithms / Sample complexity
Nonparametric Iterative Machine Teaching
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 greedy example selection by maximizing the difference between the current and target functions instead of random sampling | NIMT · Synthetic and real-world teaching experiments | loss | Helped |
| Use GFT-1 instead of RFT | NIMT · Cheetah impartation | convergence speed | Helped |
| Use GFT-1 instead of RFT | NIMT · Synthetic 1D Gaussian mixture | convergence to target function | Helped |
| Use GFT-1 instead of RFT | NIMT · Synthetic 2D classification | convergence to target function | Helped |
| Restrict the teacher to a pool containing 80% of pixels | GFT-1 · MNIST digit correction | convergence to target function | Hurt |
| Restrict the teacher to examples from a pool | GFT-1 · Cheetah impartation | teaching completeness | Hurt |
| Occasionally use character O as an alternative example, with probability 0.2 | GFT-1 · MNIST digit correction | teaching performance | No clear effect |
| Increase the GFT-k pack size proportion (ablation) | GFT-k · MNIST digit correction | convergence speed | Hurt |
| Use GFT instead of RFT at the same k | GFT-k · MNIST digit correction | convergence speed | Helped |
| Teach with RFT rather than providing the whole set | RFT · MNIST digit correction | performance | No clear effect |
| Use GFT rather than RFT | NIMT · Theoretical ITD under the stated assumptions | iterative teaching dimension: O(2\mathcal{L}(f^0)/\left(\tilde{\eta}\epsilon\right)) -> O(\psi(\frac{2\mathcal{L}(f^0)}{\tilde{\eta}\epsilon})) | Helped |
Closest papers by meaning
- Nonparametric Teaching for Multiple Learners (NeurIPS 2023 · similarity 0.88)
- Iterative Machine Teaching (ICML 2017 · similarity 0.88)
- Nonparametric Teaching of Implicit Neural Representations (ICML 2024 · similarity 0.84)
- Sharpness-Aware Minimization Activates the Interactive Teaching's Understanding and Optimization (NeurIPS 2024 · similarity 0.83)
- Nonparametric Teaching for Graph Property Learners (ICML 2025 · similarity 0.82)
- Smoothing Proximal Gradient Methods for Nonsmooth Sparsity Constrained Optimization: Optimality Conditions and Global Convergence (ICML 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.