Select a calendar:
Filter December Events by Event Type:
SUNMONTUEWEDTHUFRISAT
Events for December 05, 2006
-
STOP DAY
Tue, Dec 05, 2006
Viterbi School of Engineering Student Affairs
University Calendar
Audiences: Everyone Is Invited
Contact: VSOE
-
Protocols for Adaptive Modulation and Coding in Dynamic Spectrum Access Networks
Tue, Dec 05, 2006 @ 11:00 AM - 12:30 PM
Ming Hsieh Department of Electrical and Computer Engineering
Conferences, Lectures, & Seminars
SPEAKER: Michael B. Pursley, Holcombe Professor, Clemson UniversityABSTRACT: A protocol suite is presented for controlling transmissions in dynamic spectrum access networks. A framework is provided for the selection of the initial modulation to be used in a session after a frequency band has been designated. During the first few packet transmissions in a new session, a power-adjustment protocol compensates for uncertainties in the propagation characteristics and interference in the designated frequency band. Throughout the session, the error-control code and modulation are adapted to accommodate time-varying interference and propagation loss. Because increases in transmitter power can disrupt other sessions that are underway in the network, the transmitter power is increased only if adaptation of coding and modulation cannot compensate fully for deteriorations in the channel conditions. Protocol performance results are provided for static channels with unknown characteristics and for dynamic channels that are modeled as finite-state Markov chains. Comparisons with Shannon capacity limits are employed to assess the efficiency of the adaptive coding protocol for a wide range of modulation techniques and channel models.Host: Prof. Robert Scholtz, scholtz@usc.edu
Location: Hughes Aircraft Electrical Engineering Center (EEB) - 248
Audiences: Everyone Is Invited
Contact: Mayumi Thrasher
-
Tutor Days
Tue, Dec 05, 2006 @ 02:00 PM - 05:00 PM
Viterbi School of Engineering Student Affairs
University Calendar
Come to RTH to join study sessions for the following classes: EE101, ISE 460, MATH 125, MATH 126, MATH 226, PHYS 151, PHYS 152 and PHYS 153. There will be snacks, packages and upper division "study partners" there to offer support!No RSVP required...feel free to stop by any time!
Location: Ronald Tutor Hall - 1st Floor
Audiences: Viterbi Undergraduates
Contact: Viterbi Student Affairs
-
USC CS Colloquium Lecture Series
Tue, Dec 05, 2006 @ 03:30 PM - 05:00 PM
Thomas Lord Department of Computer Science
Conferences, Lectures, & Seminars
An Zhu
GoogleTitle: Towards Achieving Anonymity Abstract:We study the problem of publishing data from a table containing personal data, in a privacy preserving manner. In particular, we aim to anonymize quasi-identifiers, i.e., non-key attributes that combinedly identifiy a unique record in the table.The first model is proposed by Sweeney, called k-anonymity. This approach suppresses some values of the quasi-identifiers, such that for every record in the modified table, there are at least k-1 other records with exactly the same value. And the quality measure here is the number of quasi-identifier values suppressed. We provide a O(k)-approximation algorithm for this problem, improving upon the previous O(k log k) result. We also show that this is the best approximation bound possible using the distance representation. For small values of k, we provide improved bounds as well.We propose a second model which generalizes the quasi-identifier values via clustering. The records are first clustered and then the cluster centers are published. To ensure privacy, we impose the constraint that each cluster must contain at least k records. We consider the measure of minimizing the maximum cluster radius, for which we provide a tight 2-approximation algorithm. The second measure concerns minimizing the sum, over all clusters, the product of number of records per cluster and the cluster radius. For this measure we also provide a constant approximation algorithm. Further, we extend the algorithms to handle the case where we can omit outliners.This talk is based on two papers:
Anonymizing Tables (ICDT 05), coauthored with Aggarwal, Feder, Kenthapadi, Motwani, Panigrahy, and Thomas.
Achieveing Anonymity via Clustering (PODS 06), coauthored with Aggrawal, Feder, Kenthapadi, Khuller, Panigrahy, and Thomas.Bio:
An obtained her phd from Stanford University in 2004, under the supervision of Rajeev Motwani and Leo Guibas. An joined Google after her graduation. Since then, An has been involved in a variety of projects at Google, including search quality/ranking, image search, scholar search, and search infrastructureLocation: Seaver Science Library (SSL) - 150
Audiences: Everyone Is Invited
Contact: Nancy Levien
