Research map / Bandits and online learning / Online convex optimization and games
Neural Pseudo-Label Optimism for the Bank Loan Problem
Aldo Pacchiano, Shaun Singh, Edward Chou, Alexander C. Berg, Jakob Nicolaus Foerster
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 |
|---|---|---|---|
| train for XOR boundary: more batches seen (batch 1 to batch 81) | PLOT · XOR toy dataset | test accuracy (%): $50\%$ -> $97.8\%$ | Improved |
| more batches seen (batch 81 to batch 105) | PLOT · XOR toy dataset | test accuracy (%): $97.8\%$ -> $99.9\%$ | Improved |
| use optimistic pseudo-labels rather than NeuralUCB / greedy exploration | neural bank loan learner · MNIST (digit 5 vs rest) | cumulative regret | Improved |
| PLOT versus NeuralUCB and Greedy on run-to-run variance | PLOT · Adult, Bank, MNIST | variance of cumulative regret | Improved |
| PLOT versus NeuralUCB / epsilon-greedy / greedy on Adult and Bank | PLOT · Adult, Bank | cumulative regret | No clear effect |
| default PLOT weight 1 and radius infinity without tuning | PLOT · Adult, Bank, MNIST | cumulative regret | No clear effect |
| pseudo-label optimism instead of decayed epsilon-greedy | PLOT · Adult, Bank, MNIST | probability of accepting positives vs negatives: 3-5 times | Improved |
Closest papers by meaning
- Overcoming the Optimizer's Curse: Obtaining Realistic Prescriptions from Neural Networks (ICML 2024 · similarity 0.83)
- Arboreal Neural Network (ICML 2026 · similarity 0.82)
- The Label Horizon Paradox: Rethinking Supervision Targets in Financial Forecasting (ICML 2026 · similarity 0.82)
- Regretful Decisions under Label Noise (ICLR 2025 · similarity 0.82)
- Feature Responsiveness Scores: Model-Agnostic Explanations for Recourse (ICLR 2025 · similarity 0.81)
- Prediction without Preclusion: Recourse Verification with Reachable Sets (ICLR 2024 · similarity 0.81)
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.
