Ray Kresman

Privacy Preserving Data Sharing With Anonymous ID Assignment

An algorithm for anonymous sharing of private data among N parties is developed. This technique is used iteratively to assign these nodes ID numbers ranging from 1 to N. This assignment is anonymous in that the identities received are unknown to the other members of the group. Resistance to collusion among other members is verified in an information theoretic sense when private communication channels are used. This assignment of serial numbers allows more complex data to be shared and has applications to other problems in privacy preserving data mining, collision avoidance in communications and distributed database access. The required computations are distributed without using a trusted central authority. Existing and new algorithms for assigning anonymous IDs are examined with respect to trade-offs between communication and computational requirements. The new algorithms are built on top of a secure sum data mining operation using Newtons identities and Sturms theorem. An algorithm for distributed solution of certain polynomials over finite fields enhances the scalability of the algorithms. Markov chain representations are used to find statistics on the number of iterations required, and computer algebra gives closed form results for the completion rates.

Anonymous ID Assignment and Opt-Out

The networked society places great demand on the dissemination and sharing of private data. As privacy concerns grow, anonymity of communications becomes important. This paper addresses the issue of anonymous ID assignment to nodes in a distributed network and how it can be integrated with secure mining algorithms to allow nodes, that have privacy concerns, a capability to opt out of the mining computation. We propose two algorithms for ID assignment and evaluate their performance. We use them in the design of a protocol that allows a node to opt out of data mining, and investigate the collusion resistance capability of the resulting protocol.

Indirect Disclosures in Data Mining

Privacy preserving distributed mining algorithms mine distributed data while ensuring that ones private contribution to the global computation is not revealed. However, there are instances when such privacy assurances may fail. For example, if ones contribution happens to be an outlier, its data can be estimated from the globally mined data. In this paper we propose two simple protocols to address such indirect disclosure issues. Our work, though simple, is a bit novel: the first protocol establishes a direct relationship between a well known problem - dining cryptographers - and ours, while the second protocol extends an existing approach to computing global sum.

An Improved Algorithm for Querying Encrypted Data in the Cloud

Organizations have begun outsourcing management of their data to third party cloud service providers after the introduction of Database as a Service (DAS) model. A cloud database is a database that typically runs on a cloud computing platform, such as Amazon EC2, GoGrid, Salesforce and Rackspace. But outsourcing the data raises concerns over privacy. A typical solution is to store databases in encrypted form on the remote server. Queried records are downloaded from the server and decrypted for further processing. Bucketization is one technique for executing queries over encrypted data on a DAS server. This paper is an extension to work done by other researchers [1-4]. Query Optimal Bucketization (QOB) algorithm [1-2] divides the server data into buckets subject to an optimality constraint. In an earlier paper [3], the authors proposed Binary Query Bucketization (BQB) to improve the search time for bucketized datasets and reduce the number of records that are processed by QOB. In this paper, we propose a Parallel Binary Query Bucketization (PBQB) algorithm to query records located in the DAS. It integrates parallel search [4] and BQB. Parallel search divides the search workload into chunks with each thread/processor working on a chunk. Simulation is used to assess the numerical performance of PBQB. It is shown that the proposed algorithm outperforms BQB.

A Heuristic Algorithm for Finding Edge Disjoint Cycles in Graphs

The field of data mining provides techniques for new knowledge discovery. Distributed mining offers the miner a larger dataset with the possibility of finding stronger and, perhaps, novel association rules. This paper addresses the role of Hamiltonian cycles on mining distributed data while respecting privacy concerns. We propose a new heuristic algorithm for discovering disjoint Hamiltonian cycles. We use synthetic data to evaluate the performance of the algorithm and compare it with a greedy algorithm.

Trust Enabled Secure Multiparty Computation

Hamiltonian cycles play an important role in graph theory and data mining applications. Two Hamiltonian cycles that don’t have an edge in common are known as edge-disjoint Hamiltonian cycles (EDHCs). EDHCs are useful in computer networks. They have found applications in improving network capacity, fault-tolerance and collusion resistant mining algorithms. This paper extends previous work on collusion resistance capability of data mining algorithms. We first propose a new trust model for network computers. We then use this model as a basis to improve the collusion resistance capability of data mining algorithms. We use a performance metric to quantify the improvement.

On improving the performance of an open-source oriental game

This paper describes the design of an algorithm that improves the performance of an open-source Xiangqi (Chinese chess) software program. Our algorithm includes a new database, Opening Book Database (OB). The opening stage of a game provides for many possible moves. By importing the moves from professional players, the OB provides an efficient and fast initial move according to the position. Zobrist Hash is used to implement the OB, which makes efficient use of the storage space.

On Privacy Preserving Encrypted Data Stores

Bucketization techniques allow for effective organization of encrypted data at untrusted servers and for querying by clients. This paper presents a new metric for estimating the risk of data exposure over a set of bucketized data. The metric accounts for the importance of bucket distinctness relative to bucket access. Additionally, we review a method of controlled diffusion which improves bucket security by maximizing entropy and variance. In conjunction with our metric we use this method to show that the advantages of bucketization may be offset due to a loss of bucket security.

An Application Architecture for E-Voting

Protocols for electronic voting are becoming popular with the widespread use of internet. In this paper we propose an e-voting protocol. Our architecture is unique in that it attempts to integrate two topics of much relevance in e-voting: secure coding and information security. The former - sometimes referred to as defensive programming - provides defense against certain exploits while the latter deals with confidentiality, integrity and availability of information.