🏘️ ProcTHOR: Large-Scale Embodied AI Using Procedural Generation

Authors: Matt Deitke, Eli VanderBilt, Alvaro Herrasti, Luca Weihs, Kiana Ehsani, Jordi Salvador, Winson Han, Eric Kolve, Aniruddha Kembhavi, Roozbeh Mottaghi

NeurIPS 2022 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental We demonstrate the power and potential of PROCTHOR via a sample of 10,000 generated houses and a simple neural model. Models trained using only RGB images on PROCTHOR, with no explicit mapping and no human task supervision produce state-of-the-art results across 6 embodied AI benchmarks for navigation, rearrangement, and arm manipulation...
Researcher Affiliation Collaboration PRIOR @ Allen Institute for AI, ψUniversity of Washington, Seattle
Pseudocode No The paper describes the procedural generation process through text and diagrams (e.g., Figure 2), but does not include explicit pseudocode or algorithm blocks.
Open Source Code Yes PROCTHOR will be open-sourced and the code used in this work will be released.
Open Datasets Yes We demonstrate the power and potential of PROCTHOR using a sampled set of 10,000 fully interactive houses obtained by the procedural generation process described in Section 3 which we label PROCTHOR-10K. An additional set of 1,000 validation and 1,000 testing houses are available for evaluation.
Dataset Splits Yes An additional set of 1,000 validation and 1,000 testing houses are available for evaluation. Asset splits across train/val/test are detailed in the Appendix.
Hardware Specification Yes This set of 10K houses was generated in 1 hour on a local workstation with 4 NVIDIA RTX A5000 GPUs. Experiments were run on a server with 8 NVIDIA Quadro RTX 8000 GPUs.
Software Dependencies No The paper lists numerous open-source packages used (e.g., 'Py Torch [41]', 'Num Py [23]', 'Tensor Flow [1]') but does not provide specific version numbers for these software dependencies, only citations to their original papers or general project pages.
Experiment Setup Yes All models are trained with the Allen Act [56] framework, see the Appendix for training details.