Adaptive Coordination in Social Embodied Rearrangement

Andrew Szot, Unnat Jain, Dhruv Batra, Zsolt Kira, Ruta Desai, Akshara Rai
Proceedings of the 40th International Conference on Machine Learning, PMLR 202:33365-33380, 2023.

Abstract

We present the task of "Social Rearrangement", consisting of cooperative everyday tasks like setting up the dinner table, tidying a house or unpacking groceries in a simulated multi-agent environment. In Social Rearrangement, two robots coordinate to complete a long-horizon task, using onboard sensing and egocentric observations, and no privileged information about the environment. We study zero-shot coordination (ZSC) in this task, where an agent collaborates with a new partner, emulating a scenario where a robot collaborates with a new human partner. Prior ZSC approaches struggle to generalize in our complex and visually rich setting, and on further analysis, we find that they fail to generate diverse coordination behaviors at training time. To counter this, we propose Behavior Diversity Play (BDP), a novel ZSC approach that encourages diversity through a discriminability objective. Our results demonstrate that BDP learns adaptive agents that can tackle visual coordination, and zero-shot generalize to new partners in unseen environments, achieving 35% higher success and 32% higher efficiency compared to baselines.

Cite this Paper


BibTeX
@InProceedings{pmlr-v202-szot23a, title = {Adaptive Coordination in Social Embodied Rearrangement}, author = {Szot, Andrew and Jain, Unnat and Batra, Dhruv and Kira, Zsolt and Desai, Ruta and Rai, Akshara}, booktitle = {Proceedings of the 40th International Conference on Machine Learning}, pages = {33365--33380}, year = {2023}, editor = {Krause, Andreas and Brunskill, Emma and Cho, Kyunghyun and Engelhardt, Barbara and Sabato, Sivan and Scarlett, Jonathan}, volume = {202}, series = {Proceedings of Machine Learning Research}, month = {23--29 Jul}, publisher = {PMLR}, pdf = {https://proceedings.mlr.press/v202/szot23a/szot23a.pdf}, url = {https://proceedings.mlr.press/v202/szot23a.html}, abstract = {We present the task of "Social Rearrangement", consisting of cooperative everyday tasks like setting up the dinner table, tidying a house or unpacking groceries in a simulated multi-agent environment. In Social Rearrangement, two robots coordinate to complete a long-horizon task, using onboard sensing and egocentric observations, and no privileged information about the environment. We study zero-shot coordination (ZSC) in this task, where an agent collaborates with a new partner, emulating a scenario where a robot collaborates with a new human partner. Prior ZSC approaches struggle to generalize in our complex and visually rich setting, and on further analysis, we find that they fail to generate diverse coordination behaviors at training time. To counter this, we propose Behavior Diversity Play (BDP), a novel ZSC approach that encourages diversity through a discriminability objective. Our results demonstrate that BDP learns adaptive agents that can tackle visual coordination, and zero-shot generalize to new partners in unseen environments, achieving 35% higher success and 32% higher efficiency compared to baselines.} }
Endnote
%0 Conference Paper %T Adaptive Coordination in Social Embodied Rearrangement %A Andrew Szot %A Unnat Jain %A Dhruv Batra %A Zsolt Kira %A Ruta Desai %A Akshara Rai %B Proceedings of the 40th International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2023 %E Andreas Krause %E Emma Brunskill %E Kyunghyun Cho %E Barbara Engelhardt %E Sivan Sabato %E Jonathan Scarlett %F pmlr-v202-szot23a %I PMLR %P 33365--33380 %U https://proceedings.mlr.press/v202/szot23a.html %V 202 %X We present the task of "Social Rearrangement", consisting of cooperative everyday tasks like setting up the dinner table, tidying a house or unpacking groceries in a simulated multi-agent environment. In Social Rearrangement, two robots coordinate to complete a long-horizon task, using onboard sensing and egocentric observations, and no privileged information about the environment. We study zero-shot coordination (ZSC) in this task, where an agent collaborates with a new partner, emulating a scenario where a robot collaborates with a new human partner. Prior ZSC approaches struggle to generalize in our complex and visually rich setting, and on further analysis, we find that they fail to generate diverse coordination behaviors at training time. To counter this, we propose Behavior Diversity Play (BDP), a novel ZSC approach that encourages diversity through a discriminability objective. Our results demonstrate that BDP learns adaptive agents that can tackle visual coordination, and zero-shot generalize to new partners in unseen environments, achieving 35% higher success and 32% higher efficiency compared to baselines.
APA
Szot, A., Jain, U., Batra, D., Kira, Z., Desai, R. & Rai, A.. (2023). Adaptive Coordination in Social Embodied Rearrangement. Proceedings of the 40th International Conference on Machine Learning, in Proceedings of Machine Learning Research 202:33365-33380 Available from https://proceedings.mlr.press/v202/szot23a.html.

Related Material