Scaling up and Stabilizing Differentiable Planning with Implicit Differentiation
Authors: Linfeng Zhao, Huazhe Xu, Lawson L.S. Wong
ICLR 2023 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | We study the convergence stability, scalability, and efficiency of the proposed implicit version of VIN and its variants and demonstrate their superiorities on a range of planning tasks: 2D navigation, visual navigation, and 2-DOF manipulation in configuration space and workspace. |
| Researcher Affiliation | Academia | 1Khoury College of Computer Sciences, Northeastern University, 2Institute for Interdisciplinary Information Sciences, Tsinghua University, 3Shanghai Qi Zhi Institute |
| Pseudocode | No | The paper does not contain a structured pseudocode or algorithm block. |
| Open Source Code | No | We plan to open-source our code next. |
| Open Datasets | No | The paper states that data was randomly generated (e.g., 'randomly generate 10K/2K/2K maps') and mentions using the Gym Mini World environment, but it does not provide concrete access information (e.g., specific links, DOIs, or repository names) for the generated datasets or the generated maps. |
| Dataset Splits | Yes | We randomly generate training, validation and test data of 10K/2K/2K maps for all map sizes. |
| Hardware Specification | Yes | We use RTX 2080 Ti cards with 11GB memory for training, thus we use 11GB as the memory limit for all models. |
| Software Dependencies | No | The paper mentions 'Gym Mini World (Chevalier-Boisvert, 2018)' which is a citation for an environment, but it does not list specific version numbers for software dependencies such as libraries or frameworks (e.g., PyTorch, TensorFlow, etc.) used in their implementation. |
| Experiment Setup | Yes | The training process (on given maps) follows (Tamar et al., 2016; Lee et al., 2018; Zhao et al., 2022), where we train 60 epochs with batch size 32, and use kernel size F = 3 by default... We use Klayer = 30, 50, 80 iterations for ADPs... We fix the number of iterations of backward solver as Kbwd = 15 and of forward solver as Kfwd = 30, 50, 80. |