Peter Stone's Selected Publications

Classified by TopicClassified by Publication TypeSorted by DateSorted by First Author Last NameClassified by Funding Source


Half Field Offense: An Environment for Multiagent Learning and Ad Hoc Teamwork

Half Field Offense: An Environment for Multiagent Learning and Ad Hoc Teamwork.
Matthew Hausknecht, Prannoy Mupparaju, Sandeep Subramanian, Shivaram Kalyanakrishnan, and Peter Stone.
In AAMAS Adaptive Learning Agents (ALA) Workshop, May 2016.

Download

[PDF]253.9kB  

Abstract

The RoboCup 2D simulation domain has served as a platform for researchin AI, machine learning, and multiagent systems for more than twodecades. However, for the researcher looking to quickly prototype andevaluate different algorithms, the full RoboCup task presents acumbersome prospect, as it can take several weeks to set up thedesired testing environment. The complexity owes in part to thecoordination of several agents, each with a multi-layered controlhierarchy, and which must balance offensive and defensive goals. Thispaper introduces a new open source benchmark, based on the Half FieldOffense (HFO) subtask of soccer, as an easy-to-use platform forexperimentation. While retaining the inherent challenges of soccer,the HFO environment constrains the agent's attention todecision-making, providing standardized interfaces for interactingwith the environment and with other agents, and standardized tools forevaluating performance. The resulting testbed makes it convenient totest algorithms for single and multiagent learning, ad hoc teamwork,and imitation learning. Along with a detailed description of the HFOenvironment, we present benchmark results for reinforcement learningagents on a diverse set of HFO tasks. We also highlight several otherchallenges that the HFO environment opens up for future research.

BibTeX Entry

@InProceedings{ALA16-hausknecht,
  author = {Matthew Hausknecht and Prannoy Mupparaju and Sandeep Subramanian and Shivaram Kalyanakrishnan and Peter Stone},
  title = {Half Field Offense: An Environment for Multiagent Learning and Ad Hoc Teamwork},
  booktitle = {AAMAS Adaptive Learning Agents (ALA) Workshop},
  location = {Singapore},
  month = {May},
  year = {2016},
  abstract = {
The RoboCup 2D simulation domain has served as a platform for research
in AI, machine learning, and multiagent systems for more than two
decades. However, for the researcher looking to quickly prototype and
evaluate different algorithms, the full RoboCup task presents a
cumbersome prospect, as it can take several weeks to set up the
desired testing environment. The complexity owes in part to the
coordination of several agents, each with a multi-layered control
hierarchy, and which must balance offensive and defensive goals. This
paper introduces a new open source benchmark, based on the Half Field
Offense (HFO) subtask of soccer, as an easy-to-use platform for
experimentation. While retaining the inherent challenges of soccer,
the HFO environment constrains the agent's attention to
decision-making, providing standardized interfaces for interacting
with the environment and with other agents, and standardized tools for
evaluating performance. The resulting testbed makes it convenient to
test algorithms for single and multiagent learning, ad hoc teamwork,
and imitation learning. Along with a detailed description of the HFO
environment, we present benchmark results for reinforcement learning
agents on a diverse set of HFO tasks. We also highlight several other
challenges that the HFO environment opens up for future research.
  },
}

Generated by bib2html.pl (written by Patrick Riley ) on Mon Mar 25, 2024 00:05:18