Creating grounded and disentangled representation of robot skills
Title: Creating grounded and disentangled representation of robot skills
SNIC Project: SNIC 2021/22-910
Project Type: SNIC Small Compute
Principal Investigator: Alexander Dürr <alexander.durr@cs.lth.se>
Affiliation: Lunds universitet
Duration: 2021-11-17 – 2022-12-01
Classification: 10207
Homepage: https://portal.research.lu.se/portal/en/projects/reinforcement-learning-in-continuous-spaces-with-interactively-acquired-knowledgebased-models(6bcfa32c-e9c7-468a-9acb-f433ad98a06a).html
Keywords:

Abstract

We want to get familiar with the Tetralith resources as we attend(ed) the Singularity workshop on November 17. We want to explore the possibility to use Tetralith's setup: Singularity, GPUs, ThinLink. Our setup involves ROS and parallel simulations (-> Singularity) for which we sometimes would like to have GUI access (->ThinLink). We want to test a small scale Neural Architecture search (->GPUs). Long term, we want to develop our setup on Tetralith to ease a possible later move to Berzelius if the model scales better with more GPU power.