Advertisement

Active Learning Is Planning: Nonmyopic ε-Bayes-Optimal Active Learning of Gaussian Processes

  • Trong Nghia Hoang
  • Kian Hsiang Low
  • Patrick Jaillet
  • Mohan Kankanhalli
Part of the Lecture Notes in Computer Science book series (LNCS, volume 8726)

Abstract

A fundamental issue in active learning of Gaussian processes is that of the exploration-exploitation trade-off. This paper presents a novel nonmyopic ε-Bayes-optimal active learning (ε-BAL) approach [4] that jointly optimizes the trade-off. In contrast, existing works have primarily developed greedy algorithms or performed exploration and exploitation separately. To perform active learning in real time, we then propose an anytime algorithm [4] based on ε-BAL with performance guarantee and empirically demonstrate using a real-world dataset that, with limited budget, it outperforms the state-of-the-art algorithms.

Keywords

Gaussian Process Path Planning Road Segment Joint Entropy Gaussian Process Model 
These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

References

  1. 1.
    Cao, N., Low, K.H., Dolan, J.M.: Multi-robot informative path planning for active sensing of environmental phenomena: A tale of two algorithms. In: Proc. AAMAS, pp. 7–14 (2013)Google Scholar
  2. 2.
    Chen, J., Low, K.H., Tan, C.K.Y.: Gaussian process-based decentralized data fusion and active sensing for mobility-on-demand system. In: Proc. RSS (2013)Google Scholar
  3. 3.
    Chen, J., Low, K.H., Tan, C.K.Y., Oran, A., Jaillet, P., Dolan, J.M., Sukhatme, G.S.: Decentralized data fusion and active sensing with mobile sensors for modeling and predicting spatiotemporal traffic phenomena. In: Proc. UAI. pp. 163–173 (2012)Google Scholar
  4. 4.
    Hoang, T.N., Low, K.H., Jaillet, P., Kankanhalli, M.: Nonmyopic ε-Bayes-Optimal Active Learning of Gaussian Processes. In: Proc. ICML, pp. 739–747 (2014)Google Scholar
  5. 5.
    Krause, A., Guestrin, C.: Nonmyopic active learning of Gaussian processes: An exploration-exploitation approach. In: Proc. ICML, pp. 449–456 (2007)Google Scholar
  6. 6.
    Low, K.H., Dolan, J.M., Khosla, P.: Adaptive multi-robot wide-area exploration and mapping. In: Proc. AAMAS, pp. 23–30 (2008)Google Scholar
  7. 7.
    Low, K.H., Dolan, J.M., Khosla, P.: Information-theoretic approach to efficient adaptive path planning for mobile robotic environmental sensing. In: Proc. ICAPS (2009)Google Scholar
  8. 8.
    Low, K.H., Dolan, J.M., Khosla, P.: Active Markov information-theoretic path planning for robotic environmental sensing. In: Proc. AAMAS, pp. 753–760 (2011)Google Scholar
  9. 9.
    Martin, R.J.: Comparing and contrasting some environmental and experimental design problems. Environmetrics 12(3), 303–317 (2001)CrossRefGoogle Scholar
  10. 10.
    Ouyang, R., Low, K.H., Chen, J., Jaillet, P.: Multi-robot active sensing of non-stationary Gaussian process-based environmental phenomena. In: Proc. AAMAS, pp. 573–580 (2014)Google Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2014

Authors and Affiliations

  • Trong Nghia Hoang
    • 1
  • Kian Hsiang Low
    • 1
  • Patrick Jaillet
    • 2
  • Mohan Kankanhalli
    • 1
  1. 1.National University of SingaporeSingapore
  2. 2.Massachusetts Institute of TechnologyUSA

Personalised recommendations