# policy training Projects at AI Tinkerers

> Canonical HTML: https://aitinkerers.org/technologies/policy-training
> Markdown URL: https://aitinkerers.org/technologies/policy-training.md
> Technology record last updated: 2026-06-27T04:42:53Z
> Generated: 2026-10-24T23:15:20Z

Policy training optimizes an AI agent's decision-making strategy by using a reward model to align its outputs with human preferences.

In machine learning, policy training is the critical process of optimizing an agent's decision-making function (its policy) to maximize cumulative rewards. While traditional reinforcement learning relies on hardcoded reward formulas, modern applications utilize Reinforcement Learning from Human Feedback (RLHF) to train a separate reward model based on human rankings. This reward model then guides optimization algorithms (such as Proximal Policy Optimization, or PPO) to fine-tune the agent's policy. This approach allows developers to align complex generative models, like large language models, with nuanced human values, safety guidelines, and conversational preferences without requiring manually defined mathematical rules.

- Official technology site: https://wikipedia.org/wiki/Reinforcement_learning_from_human_feedback
- Public AI Tinkerers demos and talks: 1
- Result page: 1 of 1

## Recent Public Talks and Demos

### [Teaching a robot to always win connect four via LeRobot](https://berlin.aitinkerers.org/talks/rsvp_HveawmqfJoQ)

- Event context: AI Tinkerers Berlin - June 17, 2026 — 2026-06-17 — Berlin
- Public talk page: https://berlin.aitinkerers.org/talks/rsvp_HveawmqfJoQ

## Related Technologies

- [Action Chunking (ACT)](https://aitinkerers.org/technologies/action-chunking-act) ([Markdown](https://aitinkerers.org/technologies/action-chunking-act.md)) — 1 public demo
- [Algorithm](https://aitinkerers.org/technologies/algorithm) ([Markdown](https://aitinkerers.org/technologies/algorithm.md)) — 2 public demos
- [Chunking](https://aitinkerers.org/technologies/chunking) ([Markdown](https://aitinkerers.org/technologies/chunking.md)) — 2 public demos
- [Hardware](https://aitinkerers.org/technologies/hardware) ([Markdown](https://aitinkerers.org/technologies/hardware.md)) — 2 public demos
- [Hugging Face](https://aitinkerers.org/technologies/hugging-face) ([Markdown](https://aitinkerers.org/technologies/hugging-face.md)) — 42 public demos
- [Imitation Learning](https://aitinkerers.org/technologies/imitation-learning) ([Markdown](https://aitinkerers.org/technologies/imitation-learning.md)) — 1 public demo
- [LeRobot](https://aitinkerers.org/technologies/lerobot) ([Markdown](https://aitinkerers.org/technologies/lerobot.md)) — 1 public demo
- [Robotic Arm](https://aitinkerers.org/technologies/robotic-arm) ([Markdown](https://aitinkerers.org/technologies/robotic-arm.md)) — 1 public demo
- [SO101](https://aitinkerers.org/technologies/so101) ([Markdown](https://aitinkerers.org/technologies/so101.md)) — 1 public demo
- [Transformers](https://aitinkerers.org/technologies/transformers) ([Markdown](https://aitinkerers.org/technologies/transformers.md)) — 149 public demos
- [Transformers (ACT)](https://aitinkerers.org/technologies/transformers-act) ([Markdown](https://aitinkerers.org/technologies/transformers-act.md)) — 1 public demo