Research map / Reinforcement learning / Imitation learning
Smooth Imitation Learning for Online Sequence Prediction
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 |
|---|---|---|---|
| regularize regression-forest policy with linear autoregressor smooth class H (smooth policy class) | SIMILE policy class · automated camera planning | trajectory smoothness and accuracy (qualitative) | Helped |
| adaptive interpolation parameter beta (from empirical errors) instead of fixed conservative beta (ablation) | SIMILE · automated camera planning | convergence rate | Helped |
| deterministic policy interpolation instead of stochastic interpolation (SEARN-style) (ablation) | SIMILE · automated camera planning | imitation error | Helped |
| gradually decrease smooth feedback weight sigma toward 0 over iterations (smooth virtual feedback early) (ablation) | SIMILE feedback generation · automated camera planning | smoothness and imitation loss trade-off | Helped |
| reduce ~1 bit of data aggregation (no DAgger-style data aggregation) | SIMILE · automated camera planning | training time | No clear effect |
Closest papers by meaning
- OLLIE: Imitation Learning from Offline Pretraining to Online Finetuning (ICML 2024 · similarity 0.89)
- Is Behavior Cloning All You Need? Understanding Horizon in Imitation Learning (NeurIPS 2024 · similarity 0.89)
- Interactive and Hybrid Imitation Learning: Provably Beating Behavior Cloning (NeurIPS 2025 · similarity 0.89)
- Reward-free World Models for Online Imitation Learning (ICML 2025 · similarity 0.89)
- Policy Decorator: Model-Agnostic Online Refinement for Large Policy Model (ICLR 2025 · similarity 0.88)
- Behavior Cloning is Not All You Need: The Optimality of On-Policy Distillation for Noisy Expert Feedback (NeurIPS 2026 · similarity 0.88)
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.