Peter Stone's Selected Publications

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


A Stitch in Time - Autonomous Model Management via Reinforcement Learning

Elad Liebman, Eric Zavesky, and Peter Stone. A Stitch in Time - Autonomous Model Management via Reinforcement Learning. In Proceedings of the 17th International Conference on Autonomous Agents and Multiagent Systems (AAMAS), July 2018.

Download

[PDF]1.7MB  

Abstract

Concept drift - a change, either sudden or gradual, in the underlyingproperties of data - is one of the most prevalent challenges tomaintaining high-performing learned models over time in autonomoussystems. In the face of concept drift, one can hope that the oldmodel is sufficiently representative of the new data despite theconcept drift, one can discard the old data and retrain a new modelwith (often limited) new data, or one can use transfer learningmethods to combine the old data with the new to create an updatedmodel. Which of these three options is chosen affects not onlynear-term decisions, but also future needs to transfer or retrain. Inthis paper, we thus model response to concept drift as a sequentialdecision making problem and formally frame it as a Markov DecisionProcess. Our reinforcement learning approach to the problem showspromising results on one synthetic and two real-world datasets.

BibTeX Entry

@InProceedings{AAMAS2018-eladlieb,
  author = {Elad Liebman and Eric Zavesky and Peter Stone},
  title = {{A} {S}titch in {T}ime - {A}utonomous {M}odel {M}anagement via {R}einforcement {L}earning},
  booktitle = {Proceedings of the 17th International Conference on Autonomous Agents and Multiagent Systems (AAMAS)},
  location = {Stockholm, Sweden},
  month = {July},
  year = {2018},
  abstract = {
Concept drift - a change, either sudden or gradual, in the underlying
properties of data - is one of the most prevalent challenges to
maintaining high-performing learned models over time in autonomous
systems.  In the face of concept drift, one can hope that the old
model is sufficiently representative of the new data despite the
concept drift, one can discard the old data and retrain a new model
with (often limited) new data, or one can use transfer learning
methods to combine the old data with the new to create an updated
model.  Which of these three options is chosen affects not only
near-term decisions, but also future needs to transfer or retrain.  In
this paper, we thus model response to concept drift as a sequential
decision making problem and formally frame it as a Markov Decision
Process.  Our reinforcement learning approach to the problem shows
promising results on one synthetic and two real-world datasets.
  },
}

Generated by bib2html.pl (written by Patrick Riley ) on Wed Nov 25, 2020 10:07:50