1 Overview
Data-driven social imitation learning is a minimally-supervised approach
to generating robot behaviors for human-robot interaction (HRI).
However, this type of learning-based approach is error-prone. This
dataset contains the experiement logs of a shopkeeper robot automated
via data-driven HRI (originally from Liu et al. (2016)). It contains 200
trials, where each trial begins when a customer (human) enters a camera
shop (setup in our lab) and ends when the customer leaves. During each
episode the human customer and robot shopkeeper interact by exchanging
greetings, asking and answering questions, and moving around the shop.
In some cases the robot makes errors.
2 Robot Shopkeeper Errors in Camera Shop Customer Service Interactions
The dataset contains human labels of robot errors (from Ravishankar et
al. (2026)), including error type and how the customer responded to the
error. Please see the README.txt for more information.
2.1 File Descriptions
- HRI_data.csv – contains the HRI experiment log data and the error,
error type, and customer response type labels. Each row contains data on
a single ‘turn’ with an input (customer action) and an output
(shopkeeper action). - README.txt – contains the definitions of the fields in HRI_data.csv
and the definitions of the error, error type, and customer response type
labels.
3 Usage
The dataset is free to use for research purposes only. In case you use
the data in your work please be sure to cite the reference paper below.
4 Reference
Ravishankar, J., Doering, M., & Kanda, T., Zero-Shot Learning to Enable
Error Awareness in Data-Driven HRI, in Proceedings of the 2024 ACM/IEEE
International Conference on Human-Robot Interaction (pp. 592-601),
March, 2024. Doi: 10.1145/3610977.3634940
5 Related Work
- Liu, P., Glas, D. F., Kanda, T., & Ishiguro, H., Data-Driven HRI:
Learning Social Behaviors by Example from Human-Human Interaction, in
IEEE Transactions on Robotics, Vol. 32, No. 4, pp. 988-1008, 2016.
doi:10.1109/tro.2016.2588880 - Ravishankar, J., Doering, M., & Kanda, T., Online Continual Learning
for Automatic Error Prevention in Data-Driven HRI, in ACM Transactions
on Human-Robot Interaction, 15, 1, Article 13 (January 2026), 18 pages.
https://doi.org/10.1145/3759160