technique

Imitation Learning

techniqueactiveimitation-learning-665c03c7·1 events·first seen 28d ago

Aliases: Imitation Learning

Co-occurring entities

One-Shot Imitation Learning meta-learning OpenAI

More like this (12)

One-Shot Imitation Learning Q-learning self-play reinforcement learning unsupervised learning Reasoning Imitation Soft Q-Learning Reinforcement Learning meta-learning soft-label learning Difference-Aware Retrieval Policies for Imitation Learning Reward Learning from Comparisons few-shot learning

Recent events (1)

4Openai Blog·28d ago·source ↗

One-shot imitation learning

OpenAI published research on one-shot imitation learning, a technique enabling agents to learn new tasks from a single demonstration. The approach allows a policy network to observe a demonstration and immediately generalize to new instances of the same task without additional training. This was an early contribution to the field of meta-learning and few-shot generalization in robotics and sequential decision-making.

Agent and Tool Ecosystem Alignment and RLHF One-Shot Imitation Learning meta-learning Imitation Learning +1 more