Post-Training, Finetuning, and Steering:A Field Guide for Curating Training Data
cs.RO
Submitted: 2026-10-13
Updated: 2026-10-13
Comments: 8 pages, 6 figures
License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Terminology
Sources
- EXPO: Stable Reinforcement Learning with Expressive Policies
- Update-Free On-Policy Steering via Verifiers
- $\pi_\texttt{RL}$: Online RL Fine-tuning for Flow-based Vision-Language-Action Models
- Self-Augmented Robot Trajectory: Efficient Imitation Learning via Safe Self-augmentation with Demonstrator-annotated Precision
- Scaling Instructable Agents Across Many Simulated Worlds
- Q-learning with Adjoint Matching
- Flow Matching Policy Gradients
- Policy Agnostic RL: Offline RL and Online RL Fine-Tuning of Any Class and Backbone
- IDQL: Implicit Q-Learning as an Actor-Critic Method with Diffusion Policies
- Diffusion Guidance Is a Controllable Policy Improvement Operator
- $\pi_{0.5}$: a Vision-Language-Action Model with Open-World Generalization
- QPILOTS: Efficient Test-Time Q-Steering for Flow Policies
- RLinf: Flexible and Efficient Large-scale Reinforcement Learning via Macro-to-Micro Flow Transformation
- Meta Flow Maps enable scalable reward alignment
- DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models
Related papers
- FMT x: An Efficient and Asymptotically Optimal Extension of the Fast Marching Tree for Dynamic Replanning
- MPCFormer: A physics-informed data-driven approach for explainable socially-aware autonomous driving
- RoboLab: A High-Fidelity Simulation Benchmark for Analysis of Task Generalist Policies
- ARC: A Reasoning Recipe for Robot Foundation Models
- HRDexDB: A 4D Dexterous Grasping Dataset Across Human and Multiple Robot Embodiments
- APT: Action Expert Pretraining Improves Instruction Generalization of Vision-Language-Action Policies