Incremental Learning of Humanoid Robot Behavior from Natural Interaction and Large Language Models
A HumanoidRobots presenter demonstrates how a humanoid robot can learn household routines through natural interaction, including requests to bring drinks and cleaning tools.
In this talk
- 1:26
Learning to pair a drink with a cup
The robot first brings juice to a square table. After the person asks for a cup too, they tell it to include one whenever it brings a drink. On the next request, it brings both items without being prompted.
- 3:05
Adapting a cleaning routine
When asked to help clean the refrigerator top, the robot brings a sponge, then fetches a ladder after the person points out it is needed. The person asks it to remember the ladder for future work on high surfaces.
- 4:21
Applying the learned routine
The person revisits the refrigerator task, and the robot brings the ladder and sponge together. The examples show how spoken feedback can add missing items to a household task routine.
Briefs summarise talks and interviews; timestamps open the video at that moment. Claims in a talk are the speaker's own.