Active diagnosis via AUC maximization: an efficient approach for multiple fault identification in large scale, noisy networks
July 2011
Pages 35 - 42
Abstract
The problem of active diagnosis arises in several applications such as disease diagnosis, and fault diagnosis in computer networks, where the goal is to rapidly identify the binary states of a set of objects (e.g., faulty or working) by sequentially selecting, and observing, (noisy) responses to binary valued queries. Current algorithms in this area rely on loopy belief propagation for active query selection. These algorithms have an exponential time complexity, making them slow and even intractable in large networks. We propose a rank-based greedy algorithm that sequentially chooses queries such that the area under the ROC curve of the rank-based output is maximized. The AUC criterion allows us to make a simplifying assumption that significantly reduces the complexity of active query selection (from exponential to near quadratic), with little or no compromise on the performance quality.
References
[1]
K. Ataman, W. N. Street, and Y. Zhang. Learning to rank by maximizing AUC with linear programming. Proc. IJCNN, pages 123-129, 2006.
[2]
G. Bellala, J. Stanley, C. Scott, and S. K. Bhavnani. Supplemental: Active diagnosis via AUC maximization. 2011. available at www-personal.umich.edu/~gowtham.
[3]
L. Cheng, X. Qui, L. Meng, Y. Qiao, and R. Boutaba. Efficient active probing for fault diagnosis in large scale and noisy networks. IEEE INFOCOM, 2010.
[4]
C. Cortes and M. Mohri. AUC optimization vs. error rate minimization. In NIPS, 2003.
[5]
C. Ferri, P. Flach, and J. Hernández-Orallo. Learning decision trees using the area under the ROC curve. In ICML, 2002.
[6]
D. Golovin and A. Krause. Adaptive Submodularity: A new approach to active learning and stochastic optimization. In COLT, 2010.
[7]
J. Guillaume and M. Latapy. Bipartite Graphs as Models of Complex Networks. Springer, 2004.
[8]
D. Heckerman. A tractable inference algorithm for diagnosing multiple diseases. UAI, 5:163-171, 1990.
[9]
T. S. Jaakkola and M. I. Jordan. Variational methods and the QMR-DT databases. Journal of Artificial Intelligence Research, 10:291-322, 1999.
[10]
S. Kandula, D. Katabi, and J. P. Vasseur. Shrink: A tool for failure diagnosis in IP networks. Proc. ACM SIG-COMM MineNet Workshop, August 2005.
[11]
T. Le and C. N. Hadjicostis. Max-product algorithms for the generalized multiple-fault diagnosis problem. IEEE Trans. on Sys., Man and Cybern., 37(6), 2007.
[12]
P. M. Long and R. A. Servedio. Boosting the Area under the ROC curve. In NIPS, 2007.
[13]
A. Medina, A. Lakhina, I. Matta, and J. Byers. BRITE: An Approach to Universal Topology Generation. Proc. MASCOT, 2001.
[14]
J. M. Mooij. libDAI: A free and open source C++ library for discrete approximate inference in graphical models. JMLR, 2010.
[15]
K. P. Murphy, Y. Weiss, and M. Jordan. Loopy belief propagation for approximate inference: An empirical study. In UAI, pages 467-475, 1999.
[16]
J. Pearl. Probabilistic Reasoning in Intelligent Systems: Networks of Plausible Inference. Morgan Kauf., 1988.
[17]
I. Rish, M. Brodie, S. Ma, N. Odintsova, A. Beygelzimer, G. Grabarnik, and K. Hernandez. Adaptive diagnosis in distributed systems. IEEE Trans. on Neural Networks, 16(5):1088-1109, 2005.
[18]
N. I. Santoso, C. Darken, G. Povh, and J. Erdmann. Nuclear plant fault diagnosis using probabilistic reasoning. Proc. IEEE/PES, 2:714-719, 1999.
[19]
J. Winick and S. Jamin. Inet-3.0: Internet topology generator. Technical Report CSE-TR-456-02, University of Michigan, 2002.
[20]
Z. Yongli, H. Limin, and L. Jinling. Bayesian networks based approach for power systems fault diagnosis. IEEE Trans. on P. Del., 21(2):634-639, 2006.
[21]
A. X. Zheng, I. Rish, and A. Beygelzimer. Efficient test selection in active diagnosis via entropy approximation. In UAI, 2005.
Recommendations
Detector design for active fault diagnosis in closed‐loop systems
SummaryFault diagnosis of closed‐loop systems is extremely relevant for high‐precision equipment and safety critical systems. Fault diagnosis is usually divided into 2 schemes: active and passive fault diagnosis. Recent studies have highlighted some ...
A new input design framework for asymptotic active fault diagnosis with application to integrated diagnosis and control
AbstractThis paper proposes a new input design framework for set-based asymptotic active fault diagnosis and then applies it to integrated design of active fault diagnosis and control. First, a new input design method is proposed for asymptotic active ...
Manifold optimal experimental design via dependence maximization for active learning
Naturally occurring data have been growing in a huge volume size, which poses a big challenge to give them high-quality labels to learn a good model. Therefore, it is critical to only select the most informative data points for labeling, which is cast ...
Comments
Information & Contributors
Information
Published In
Sponsors
- Pascal Network of Excellence: Pascal Network of Excellence
- Google Inc.
- Artificial Intelligence Journal
- IBMR: IBM Research
- Microsoft Research: Microsoft Research
Publisher
AUAI Press
Arlington, Virginia, United States
Publication History
Published: 14 July 2011
Qualifiers
- Article
Contributors
Other Metrics
Bibliometrics & Citations
Bibliometrics
Article Metrics
- 0Total Citations
- 0Total Downloads
- Downloads (Last 12 months)0
- Downloads (Last 6 weeks)0