Behavior Transformers: Cloning $k$ modes with one stone

Authors: Nur Muhammad Shafiullah, Zichen Cui, Ariuntuya (Arty) Altanzaya, Lerrel Pinto

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

Reproducibility Variable Result LLM Response
Research Type Experimental We experimentally evaluate Be T on a variety of robotic manipulation and self-driving behavior datasets. We show that Be T significantly improves over prior state-of-the-art work on solving demonstrated tasks while capturing the major modes present in the pre-collected datasets. Finally, through an extensive ablation study, we analyze the importance of every crucial component in Be T.
Researcher Affiliation Academia Nur Muhammad (Mahi) Shafillah Zichen Jeff Cui Ariuntuya Altanzaya Lerrel Pinto New York University Corresponding author, email: mahi@cs.nyu.edu
Pseudocode No The paper includes diagrams illustrating the architecture and process, but no structured pseudocode or algorithm blocks.
Open Source Code Yes All of our datasets, code, and trained models will be made publicly available.
Open Datasets Yes CARLA [21] uses the Unreal Engine to provide a simulated driving environment in a visually realistic landscape. [21] A. Dosovitskiy, G. Ros, F. Codevilla, A. M. Lopez, and V. Koltun. Carla: An open urban driving simulator. In Conference on robot learning, pages 1-16. PMLR, 2017.
Dataset Splits No The paper does not provide specific percentages or counts for training/validation/test splits, nor does it reference predefined splits with citations for dataset partitioning.
Hardware Specification Yes Our models contain on the order of 10^4-10^6 parameters, and even with a small batch size trains within an hour for our largest datasets (Block push) on a single desktop GPU.
Software Dependencies No The paper mentions various models and techniques but does not provide specific software environment details or library versions (e.g., Python version, PyTorch version).
Experiment Setup No The paper describes the model architecture and loss functions, but it does not provide specific hyperparameter values (e.g., learning rate, batch size, number of epochs) or detailed system-level training settings in the main text.