notesum.ai
Published at November 19Preference-Conditioned Gradient Variations for Multi-Objective Quality-Diversity
cs.AI
Released Date: November 19, 2024
Authors: Hannah Janmohamed1, Maxence Faldor2, Thomas Pierrot1, Antoine Cully2
Aff.: 1InstaDeep; 2Adaptive and Intelligent Robotics Lab, Imperial College London
![[Uncaptioned image]](https://arxiv.org/html/2411.12433v1/extracted/6010151/Experiments/figures/ant.png){kind=small}
|
![[Uncaptioned image]](https://arxiv.org/html/2411.12433v1/extracted/6010151/Experiments/figures/halfcheetah.png){kind=small}
|
![[Uncaptioned image]](https://arxiv.org/html/2411.12433v1/extracted/6010151/Experiments/figures/hopper.png){kind=small}
|
![[Uncaptioned image]](https://arxiv.org/html/2411.12433v1/extracted/6010151/Experiments/figures/walker2d.png){kind=small}
|
|||
| Name | Ant-2 | Ant-3 | HalfCheetah-2 | Hopper-2 | Hopper-3 | Walker-2 |
| feature | Feet Contact Proportion | |||||
| Rewards |
•
Forward
velocity • Energy consumption |
• velocity • velocity • Energy consumption |
•
Forward
velocity • Energy consumption |
•
Forward
velocity • Energy consumption |
•
Forward
velocity • Jumping height • Energy consumption |
•
Forward
velocity • Energy consumption |