### Abstract

The discrete CEO Problem is considered when the agents are under Byzantine attack. That is, a malicious intruder has captured an unknown subset of the agents and reprogrammed them to increase the probability of error. Two traitor models are considered, depending on whether the traitors are able to see honest agents' messages before choosing their own. If they can, bounds are given on the error exponent with respect to the sum-rate as a function of the fraction of agents that are traitors. The number of traitors is assumed to be known to the CEO, but not their identity. If they are not able to see the honest agents' messages, an exact but uncomputable characterization of the error exponent is given. It is shown that for a given sum-rate, the minimum achievable probability of error is within a factor of two of a quantity based on the traitors simulating a false distribution to generate messages they send to the CEO. This false distribution is chosen by the traitors to increase the probability of error as much as possible without revealing their identities to the CEO. Because this quantity is always within a constant factor of the probability of error, it gives the error exponent directly.

Original language | English (US) |
---|---|

Title of host publication | 46th Annual Allerton Conference on Communication, Control, and Computing |

Pages | 1207-1214 |

Number of pages | 8 |

DOIs | |

State | Published - 2008 |

Externally published | Yes |

Event | 46th Annual Allerton Conference on Communication, Control, and Computing - Monticello, IL, United States Duration: Sep 24 2008 → Sep 26 2008 |

### Other

Other | 46th Annual Allerton Conference on Communication, Control, and Computing |
---|---|

Country | United States |

City | Monticello, IL |

Period | 9/24/08 → 9/26/08 |

### Keywords

- Byzantine attack
- Distributed source coding
- Network security
- Sensor fusion

### ASJC Scopus subject areas

- Computer Networks and Communications
- Software
- Control and Systems Engineering
- Communication

### Cite this

*46th Annual Allerton Conference on Communication, Control, and Computing*(pp. 1207-1214). [4797697] https://doi.org/10.1109/ALLERTON.2008.4797697

**A characterization of the error exponent for the Byzantine CEO Problem.** / Kosut, Oliver; Tong, Lang.

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

*46th Annual Allerton Conference on Communication, Control, and Computing.*, 4797697, pp. 1207-1214, 46th Annual Allerton Conference on Communication, Control, and Computing, Monticello, IL, United States, 9/24/08. https://doi.org/10.1109/ALLERTON.2008.4797697

}

TY - GEN

T1 - A characterization of the error exponent for the Byzantine CEO Problem

AU - Kosut, Oliver

AU - Tong, Lang

PY - 2008

Y1 - 2008

N2 - The discrete CEO Problem is considered when the agents are under Byzantine attack. That is, a malicious intruder has captured an unknown subset of the agents and reprogrammed them to increase the probability of error. Two traitor models are considered, depending on whether the traitors are able to see honest agents' messages before choosing their own. If they can, bounds are given on the error exponent with respect to the sum-rate as a function of the fraction of agents that are traitors. The number of traitors is assumed to be known to the CEO, but not their identity. If they are not able to see the honest agents' messages, an exact but uncomputable characterization of the error exponent is given. It is shown that for a given sum-rate, the minimum achievable probability of error is within a factor of two of a quantity based on the traitors simulating a false distribution to generate messages they send to the CEO. This false distribution is chosen by the traitors to increase the probability of error as much as possible without revealing their identities to the CEO. Because this quantity is always within a constant factor of the probability of error, it gives the error exponent directly.

AB - The discrete CEO Problem is considered when the agents are under Byzantine attack. That is, a malicious intruder has captured an unknown subset of the agents and reprogrammed them to increase the probability of error. Two traitor models are considered, depending on whether the traitors are able to see honest agents' messages before choosing their own. If they can, bounds are given on the error exponent with respect to the sum-rate as a function of the fraction of agents that are traitors. The number of traitors is assumed to be known to the CEO, but not their identity. If they are not able to see the honest agents' messages, an exact but uncomputable characterization of the error exponent is given. It is shown that for a given sum-rate, the minimum achievable probability of error is within a factor of two of a quantity based on the traitors simulating a false distribution to generate messages they send to the CEO. This false distribution is chosen by the traitors to increase the probability of error as much as possible without revealing their identities to the CEO. Because this quantity is always within a constant factor of the probability of error, it gives the error exponent directly.

KW - Byzantine attack

KW - Distributed source coding

KW - Network security

KW - Sensor fusion

UR - http://www.scopus.com/inward/record.url?scp=64549154459&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=64549154459&partnerID=8YFLogxK

U2 - 10.1109/ALLERTON.2008.4797697

DO - 10.1109/ALLERTON.2008.4797697

M3 - Conference contribution

SN - 9781424429264

SP - 1207

EP - 1214

BT - 46th Annual Allerton Conference on Communication, Control, and Computing

ER -