Sparse Latent Space Policy Search
Authors: Kevin Luck, Joni Pajarinen, Erik Berger, Ville Kyrki, Heni Ben Amor
AAAI 2016 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | In experiments with a simulated bi-manual manipulator, the new approach effectively identifies synergies between joints, performs efficient low-dimensional policy search, and outperforms state-of-the-art policy search methods. |
| Researcher Affiliation | Academia | Kevin Sebastian Luck Arizona State University Interactive Robotics Lab AZ 85281 Tempe, USA mail@kevin-luck.net Joni Pajarinen Aalto University Intelligent Robotics Group 02150 Espoo, Finland Joni.Pajarinen@aalto.fi Erik Berger Technical University Bergakademie Freiberg Institute of Computer Science 09599 Freiberg, Germany erik.berger@informatik.tu-freiberg.de Ville Kyrki Aalto University Intelligent Robotics Group 02150 Espoo, Finland ville.kyrki@aalto.fi Heni Ben Amor Arizona State University Interactive Robotics Lab AZ 85281 Tempe, USA hbenamor@asu.edu |
| Pseudocode | Yes | Algorithm 1: Outline of the Group Factor Policy Search (Grou PS) algorithm. |
| Open Source Code | No | The paper does not provide any statement or link indicating that the source code for the described methodology is publicly available. |
| Open Datasets | No | The paper uses simulated environments (a simulated bi-manual robot, NAO robot using V-REP framework) for experiments. It does not refer to or provide access information for a publicly available or open dataset. |
| Dataset Splits | No | The paper mentions sampling trajectories and selecting the '15 best trajectories' for parameter computation, but it does not specify explicit training, validation, or test dataset splits in terms of percentages or sample counts. |
| Hardware Specification | No | The paper states that experiments were conducted in a simulated environment ('simulated bi-manual manipulator', 'NAO robot using the V-REP framework') but does not specify any hardware (GPU, CPU, memory, etc.) used for these simulations. |
| Software Dependencies | No | The paper mentions 'V-REP framework (Rohmer, Singh, and Freese 2013)' and 'bullet physics library' but does not provide specific version numbers for these or other software components. |
| Experiment Setup | Yes | The hyper-parameters of Grou PS were set to a τ = b τ = 1000, aα = bα = 1 and σ2 = 100. ... The static variance parameter for Po WER ... and the initial variance of the other algorithms were all set to 101.5, also for NAC with learning parameter set to 0.5. In each iteration, we sampled 30 trajectories and evaluated the trajectories based on the reward function ... Then the 15 best trajectories are chosen for the computation of the parameters for each algorithm. |