A Model-Agnostic Approach for Semantically Driven Disambiguation in Human-Robot Interaction

Overview

This repository contains the user experiment dataset for the "A Model-Agnostic Approach for Semantically Driven Disambiguation in Human-Robot Interaction" paper.

Paper

• Title: A Model-Agnostic Approach for Semantically Driven Disambiguation in Human-Robot Interaction
• Authors: F. I. Dogan, M. Patel, W. Liu, I. Leite, S. Chernova
• Link: https://doi.org/10.48550/arXiv.2409.17004

Contents

• Object Images: Images of described objects
• Home Videos: Home clip videos from MatterPort
• Expression Dataset: Expressions collected from MTurk workers describing objects
• Object Features Dataset: Pickle file containing a complete list of object properties
• process_data.py: An example code parsing the Object Features Dataset

Dataset Structure

├── home_videos/
├── expression_data/
├── evaluation_objects/
├── process_data.py
├── object_features_data.obj
├── README.md

Cite

@article{dogan2024semantically,
  title={A Model-Agnostic Approach for Semantically Driven Disambiguation in Human-Robot Interaction},
  author={Dogan, Fethiye Irmak and Patel, Maithili and Liu, Weiyu and Leite, Iolanda and Chernova, Sonia},
  journal={arXiv preprint arXiv:2409.17004},
  year={2024}
}

Contact

For questions, reach out to [irmakdogan92@gmail.com].