AT&T Labs

About: AT&T Labs is a based out in . It is known for research contribution in the topics: Network packet & The Internet. The organization has 1879 authors who have published 5595 publications receiving 483151 citations.

A Grasp for Job Shop Scheduling

Abstract: In this paper, we describe a greedy randomized adaptive search procedure (GRASP) for the job shop scheduling problem (JSP). We incorporate to the conventional GRASP two new concepts: an intensification strategy and POP (Proximate Optimality Principle) in the construction phase. These two concepts were first proposed by Fleurent and Glover (1999) in the context of the quadratic assignment problem. Computational experience on a large set of standard test problems indicates that GRASP is a competitive algorithm for finding approximate solutions of the job shop scheduling problem.

Do too many cooks spoil the broth? Using the number of developers to enhance defect prediction models

Abstract: Fault prediction by negative binomial regression models is shown to be effective for four large production software systems from industry. A model developed originally with data from systems with regularly scheduled releases was successfully adapted to a system without releases to identify 20% of that system's files that contained 75% of the faults. A model with a pre-specified set of variables derived from earlier research was applied to three additional systems, and proved capable of identifying averages of 81, 94 and 76% of the faults in those systems. A primary focus of this paper is to investigate the impact on predictive accuracy of using data about the number of developers who access individual code units. For each system, including the cumulative number of developers who had previously modified a file yielded no more than a modest improvement in predictive accuracy. We conclude that while many factors can "spoil the broth" (lead to the release of software with too many defects), the number of developers is not a major influence.

Internet pricing with a game theoretical approach: concepts and examples

Abstract: The basic concepts of three branches of game theory, leader-follower, cooperative, and two-person nonzero sum games, are reviewed and applied to the study of the Internet pricing issue. In particular, we emphasize that the cooperative game (also called the bargaining problem) provides an overall picture for the issue. With a simple model for Internet quality of service (QoS), we demonstrate that the leader-follower game may lead to a solution that is not Pareto optimal and in some cases may be "unfair," and that the cooperative game may provide a better solution for both the Internet service provider (ISP) and the user. The practical implication of the results is that government regulation or arbitration may be helpful. The QoS model is also applied to study the competition between two ISPs, and we find a Nash equilibrium point from which the two ISPs would not move out without cooperation. The proposed approaches can be applied to other Internet pricing problems such as the Paris Metro pricing scheme.

Active learning: theory and applications to automatic speech recognition

Abstract: We are interested in the problem of adaptive learning in the context of automatic speech recognition (ASR). In this paper, we propose an active learning algorithm for ASR. Automatic speech recognition systems are trained using human supervision to provide transcriptions of speech utterances. The goal of Active Learning is to minimize the human supervision for training acoustic and language models and to maximize the performance given the transcribed and untranscribed data. Active learning aims at reducing the number of training examples to be labeled by automatically processing the unlabeled examples, and then selecting the most informative ones with respect to a given cost function for a human to label. In this paper we describe how to estimate the confidence score for each utterance through an on-line algorithm using the lattice output of a speech recognizer. The utterance scores are filtered through the informativeness function and an optimal subset of training samples is selected. The active learning algorithm has been applied to both batch and on-line learning scheme and we have experimented with different selective sampling algorithms. Our experiments show that by using active learning the amount of labeled data needed for a given word accuracy can be reduced by more than 60% with respect to random sampling.

Single database private information retrieval implies oblivious transfer

Abstract: A Single-Database Private Information Retrieval (PIR) is a protocol that allows a user to privately retrieve from a database an entry with as small as possible communication complexity. We call a PIR protocol non-trivial if its total communication is strictly less than the size of the database. Non-trivial PIR is an important cryptographic primitive with many applications. Thus, understanding which assumptions are necessary for implementing such a primitive is an important task, although (so far) not a well-understood one. In this paper we show that any non-trivial PIR implies Oblivious Transfer, a far better understood primitive. Our result not only significantly clarifies our understanding of any non-trivial PIR protocol, but also yields the following consequences: - Any non-trivial PIR is complete for all two-party and multiparty secure computations. - There exists a communication-efficient reduction from any PIR protocol to a 1-out-of-n Oblivious Transfer protocol (also called SPIR). - There is strong evidence that the assumption of the existence of a one-way function is necessary but not sufficient for any non-trivial PIR protocol.