Learning To Simulate
Authors: Nataniel Ruiz, Samuel Schulter, Manmohan Chandraker
ICLR 2019 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | In various experiments on both toy data and real computer vision problems, Section 4 analyzes different variants of our approach and investigates interesting questions... and The experiments indicate that our approach is able to quickly identify good scene parameters ψ that compete and in some cases even outperform the actual validation set parameters for synthetic as well as real data, on computer vision problems such as object counting or semantic segmentation. |
| Researcher Affiliation | Collaboration | Nataniel Ruiz 1, , Samuel Schulter 2, Manmohan Chandraker 2,3 1Department of Computer Science, Boston University 2NEC Laboratories America 3UC San Diego |
| Pseudocode | Yes | Algorithm 1: Our approach for learning to simulate based on policy gradients. |
| Open Source Code | No | The paper mentions external tools like CARLA and Unreal Engine, but it does not state that the code for the proposed learning-to-simulate method itself is publicly available, nor does it provide a link to a code repository. |
| Open Datasets | Yes | We demonstrate the practical impact of our learning-to-simulate approach on semantic segmentation on KITTI Geiger et al. (2013) |
| Dataset Splits | Yes | We generate validation and test sets from p(x, y| ψreal). |
| Hardware Specification | No | The paper does not provide specific details about the hardware (e.g., CPU, GPU models, memory, or cloud instance types) used to run the experiments. |
| Software Dependencies | No | The paper mentions using 'CARLA (Dosovitskiy et al., 2017) with Unreal engine (Epic-Games, 2018)' but does not provide specific version numbers for these or any other software components. |
| Experiment Setup | Yes | First, we define the number of training epochs ξ of the MTM in each policy iteration as a variable. The intuition is that a reasonable reward signal may be obtained even if MTM is not trained until full convergence, thus reducing computation time significantly. Second, we define the size M of the data set generated in each policy iteration. |