Share This Author
Privacy-Preserving Ridge Regression on Hundreds of Millions of Records
- V. Nikolaenko, Udi Weinsberg, Stratis Ioannidis, M. Joye, D. Boneh, N. Taft
- Computer Science, MathematicsIEEE Symposium on Security and Privacy
- 19 May 2013
This work implements the complete system and experiments with it on real data-sets, and shows that it significantly outperforms pure implementations based only on homomorphic encryption or Yao circuits.
Structural analysis of network traffic flows
- Anukool Lakhina, K. Papagiannaki, M. Crovella, C. Diot, E. Kolaczyk, N. Taft
- Computer ScienceSIGMETRICS '04/Performance '04
- 1 June 2004
This work presents the first analysis of complete sets of OD flow time-series, taken from two different backbone networks (Abilene and Sprint-Europe) and finds that the set of OD flows has small intrinsic dimension, and shows how to use PCA to systematically decompose the structure ofOD flow timeseries into three main constituents.
Traffic matrices: balancing measurements, inference and modeling
It is shown that, using such methods, small amounts of traffic flow measurements can have significant impacts on the accuracy of traffic matrix estimation, yielding results much better than previous approaches.
Privacy-preserving matrix factorization
- V. Nikolaenko, Stratis Ioannidis, Udi Weinsberg, M. Joye, N. Taft, D. Boneh
- Computer ScienceCCS
- 4 November 2013
This work shows that a recommender can profile items without ever learning the ratings users provide, or even which items they have rated, by designing a system that performs matrix factorization, a popular method used in a variety of modern recommendation systems, through a cryptographic technique known as garbled circuits.
Combining filtering and statistical methods for anomaly detection
It is explained here how any anomaly detection method can be viewed as a problem in statistical hypothesis testing, and four different methods for analyzing residuals, two of which are new are studied and compared.
ANTIDOTE: understanding and defending against poisoning of anomaly detectors
This work proposes an antidote based on techniques from robust statistics and presents a new robust PCA-based detector that substantially reduces the effectiveness of poisoning for a variety of scenarios and indeed maintains a significantly better balance between false positives and false negatives than the original method when under attack.
The problem of synthetically generating IP traffic matrices: initial recommendations
The myth that uniform distributions can be used to randomly generate numbers for populating a traffic matrix is dispelled and it is shown that the lognormal distribution is better for this purpose as it describes well the mean rates of origin-destination flows.
Learning in a Large Function Space: Privacy-Preserving Mechanisms for SVM Learning
- Benjamin I. P. Rubinstein, P. Bartlett, Ling Huang, N. Taft
- Computer ScienceJ. Priv. Confidentiality
- 30 November 2009
This paper explores the release of Support Vector Machine (SVM) classifiers while preserving the privacy of training data and presents efficient mechanisms for finite-dimensional feature mappings and for (potentially infinite-dimensional) mappings with translation-invariant kernels.
Skilled in the Art of Being Idle: Reducing Energy Waste in Networked Systems
- S. Nedevschi, J. Chandrashekar, Junda Liu, B. Nordman, S. Ratnasamy, N. Taft
- Computer ScienceNSDI
- 22 April 2009
It is found that, although there is indeed much potential for energy savings, trivial approaches are not effective and achieving substantial savings requires a careful consideration of the tradeoffs between the proxy complexity and the idle-time functionality available to users, and that these tradeoffs vary with user environment.
GraphSC: Parallel Secure Computation Made Easy
- Kartik Nayak, X. Wang, Stratis Ioannidis, Udi Weinsberg, N. Taft, E. Shi
- Computer ScienceIEEE Symposium on Security and Privacy
- 17 May 2015
This work builds Graph SC, a framework that provides a programming paradigm that allows non-cryptography experts to write secure code, brings parallelism to such secure implementations, and meets the need for obliviousness, thereby not leaking any private information.