Publications

Refine Results

(Filters Applied) Clear All

Increasing the coherence time in a magnetically-sensitive stimulated Raman transition in 85Rb

October 14, 2014
  |
Conference Paper
Author:
Sara A. DeSavage
…
Published in:
FIO 2014: Frontiers in Optics, 14 October 2014.
Topic:
quantum
R&D area:
R&D group:

Summary

We experimentally study the Ramsey, spin echo, and CPMG pulse sequences of a magnetically sensitive transition of a cold 85Rb gas. We can increase the coherence time by up to a factor of 10 by using CPMG pulse sequences as compared to Ramsey or spin echo.
READ LESS

Summary

We experimentally study the Ramsey, spin echo, and CPMG pulse sequences of a magnetically sensitive transition of a cold 85Rb gas. We can increase the coherence time by up to a factor of 10 by using CPMG pulse sequences as compared to Ramsey or spin echo.

READ MORE
Increasing the coherence time in a magnetically-sensitive stimulated Raman transition in 85Rb

Energy efficiency benefits of subthreshold-optimized transistors for digital logic

October 6, 2014
  |
Conference Paper
Author:
Peter J. Grossmann
…
Published in:
2014 IEEE SOI-3D-Subthreshold Microelectronics Technology Unified Conf. (S3S), 6-9 October 2014.
Topic:
integrated circuit
R&D area:
R&D group:

Summary

The minimum energy point of an integrated circuit (IC) is defined as the value of the supply voltage at which the energy per operation of the circuit is minimized. Several factors influence what the value of this voltage can be, including the topology of the circuit itself, the input activity factor, and the process technology in which the circuit is implemented. For application-specific ICs (ASICs), the minimum energy point usually occurs at a subthreshold supply voltage. Advances in subthreshold circuit design now permit correct circuit operation at, or even below, the minimum energy point. Since energy consumption is proportional to the square of the supply voltage, circuit design techniques and process technology choices that reduce the minimum energy point inherently improve the energy efficiency of ICs. Previous research has shown that optimizing process technology for subthreshold operation can improve IC energy efficiency. This, coupled with the energy efficiency advantages offered by fully-depleted silicon-on-insulator (FDSOI) processes, have led to the development of a subthreshold-optimized FDSOI process at MIT Lincoln Laboratory (MITLL) called xLP (Extreme Low Power). However, to date there has not been a quantitative estimate of the energy efficiency benefit of xLP or other analagous technology for complex digital circuits. This paper will show via simulation that the xLP process technology enables energy efficiency improvements that exceed that of process scaling by one generation. Specifically, the process is shown to improve power delay product by 57% vs. the IBM 90nm low power bulk process, and by 9% vs. the IBM 65 nm low power bulk technology at 0.3V.
READ LESS

Summary

The minimum energy point of an integrated circuit (IC) is defined as the value of the supply voltage at which the energy per operation of the circuit is minimized. Several factors influence what the value of this voltage can be, including the topology of the circuit itself, the input activity...

READ MORE
Energy efficiency benefits of subthreshold-optimized transistors for digital logic

Quantitative evaluation of dynamic platform techniques as a defensive mechanism

September 17, 2014
  |
Conference Paper
Author:
Hamed Okhravi
…
Published in:
RAID 2014: 17th Int. Symp. on Research in Attacks, Intrusions, and Defenses, 17-19 September 2014.
Topic:
systems analysis
R&D area:
R&D group:

Summary

Cyber defenses based on dynamic platform techniques have been proposed as a way to make systems more resilient to attacks. These defenses change the properties of the platforms in order to make attacks more complicated. Unfortunately, little work has been done on measuring the effectiveness of these defenses. In this work, we first measure the protection provided by a dynamic platform technique on a testbed. The counter-intuitive results obtained from the testbed guide us in identifying and quantifying the major effects contributing to the protection in such a system. Based on the abstract effects, we develop a generalized model of dynamic platform techniques which can be used to quantify their effectiveness. To verify and validate out results, we simulate the generalized model and show that the testbed measurements and the simulations match with small amount of error. Finally, we enumerate a number of lessons learned in our work which can be applied to quantitative evaluation of other defensive techniques.
READ LESS

Summary

Cyber defenses based on dynamic platform techniques have been proposed as a way to make systems more resilient to attacks. These defenses change the properties of the platforms in order to make attacks more complicated. Unfortunately, little work has been done on measuring the effectiveness of these defenses. In this...

READ MORE
Quantitative evaluation of dynamic platform techniques as a defensive mechanism

Using deep belief networks for vector-based speaker recognition

September 14, 2014
  |
Conference Paper
Author:
William M. Campbell
Published in:
INTERSPEECH 2014: 15th Annual Conf. of the Int. Speech Communication Assoc., 14-18 September 2014.
Topic:
human language technology
R&D area:
R&D group:

Summary

Deep belief networks (DBNs) have become a successful approach for acoustic modeling in speech recognition. DBNs exhibit strong approximation properties, improved performance, and are parameter efficient. In this work, we propose methods for applying DBNs to speaker recognition. In contrast to prior work, our approach to DBNs for speaker recognition starts at the acoustic modeling layer. We use sparse-output DBNs trained with both unsupervised and supervised methods to generate statistics for use in standard vector-based speaker recognition methods. We show that a DBN can replace a GMM UBM in this processing. Methods, qualitative analysis, and results are given on a NIST SRE 2012 task. Overall, our results show that DBNs show competitive performance to modern approaches in an initial implementation of our framework.
READ LESS

Summary

Deep belief networks (DBNs) have become a successful approach for acoustic modeling in speech recognition. DBNs exhibit strong approximation properties, improved performance, and are parameter efficient. In this work, we propose methods for applying DBNs to speaker recognition. In contrast to prior work, our approach to DBNs for speaker recognition...

READ MORE
Using deep belief networks for vector-based speaker recognition

Talking Head Detection by Likelihood-Ratio Test(220.2 KB)

September 12, 2014
  |
Conference Paper
Author:
Carl B. Quillen
…
Published in:
Second Workshop on Speech, Language, Audio in Multimedia
Topic:
machine translation
R&D area:
R&D group:

Summary

Detecting accurately when a person whose face is visible in an audio-visual medium is the audible speaker is an enabling technology with a number of useful applications. The likelihood-ratio test formulation and feature signal processing employed here allow the use of high-dimensional feature sets in the audio and visual domain, and the approach appears to have good detection performance for AV segments as short as a few seconds.
READ LESS

Summary

Detecting accurately when a person whose face is visible in an audio-visual medium is the audible speaker is an enabling technology with a number of useful applications. The likelihood-ratio test formulation and feature signal processing employed here allow the use of high-dimensional feature sets in the audio and visual domain...

READ MORE
Talking Head Detection by Likelihood-Ratio Test

A survey of cryptographic approaches to securing big-data analytics in the cloud

September 9, 2014
  |
Conference Paper
Author:
Sophia Yakoubov
…
Published in:
HPEC 2014: IEEE Conf. on High Performance Extreme Computing, 9-11 September 2014.
Topic:
supercomputing
R&D area:

Summary

The growing demand for cloud computing motivates the need to study the security of data received, stored, processed, and transmitted by a cloud. In this paper, we present a framework for such a study. We introduce a cloud computing model that captures a rich class of big-data use-cases and allows reasoning about relevant threats and security goals. We then survey three cryptographic techniques - homomorphic encryption, verifiable computation, and multi-party computation - that can be used to achieve these goals. We describe the cryptographic techniques in the context of our cloud model and highlight the differences in performance cost associated with each.
READ LESS

Summary

The growing demand for cloud computing motivates the need to study the security of data received, stored, processed, and transmitted by a cloud. In this paper, we present a framework for such a study. We introduce a cloud computing model that captures a rich class of big-data use-cases and allows...

READ MORE
A survey of cryptographic approaches to securing big-data analytics in the cloud

Computing on masked data: a high performance method for improving big data veracity

September 9, 2014
  |
Conference Paper
Author:
Jeremy Kepner
…
Published in:
HPEC 2014: IEEE Conf. on High Performance Extreme Computing, 9-11 September 2014.
Topic:
supercomputing
R&D area:
R&D group:

Summary

The growing gap between data and users calls for innovative tools that address the challenges faced by big data volume, velocity and variety. Along with these standard three V's of big data, an emerging fourth "V" is veracity, which addresses the confidentiality, integrity, and availability of the data. Traditional cryptographic techniques that ensure the veracity of data can have overheads that are too large to apply to big data. This work introduces a new technique called Computing on Masked Data (CMD), which improves data veracity by allowing computations to be performed directly on masked data and ensuring that only authorized recipients can unmask the data. Using the sparse linear algebra of associative arrays, CMD can be performed with significantly less overhead than other approaches while still supporting a wide range of linear algebraic operations on the masked data. Databases with strong support of sparse operations, such as SciDB or Apache Accumulo, are ideally suited to this technique. Examples are shown for the application of CMD to a complex DNA matching algorithm and to database operations over social media data.
READ LESS

Summary

The growing gap between data and users calls for innovative tools that address the challenges faced by big data volume, velocity and variety. Along with these standard three V's of big data, an emerging fourth "V" is veracity, which addresses the confidentiality, integrity, and availability of the data. Traditional cryptographic...

READ MORE
Computing on masked data: a high performance method for improving big data veracity

Sparse matrix partitioning for parallel eigenanalysis of large static and dynamic graphs

September 9, 2014
  |
Conference Paper
Author:
Michael M. Wolf
…
Published in:
HPEC 2014: IEEE Conf. on High Performance Extreme Computing, 9-11 September 2014.
Topic:
big data
R&D area:
R&D group:

Summary

Numerous applications focus on the analysis of entities and the connections between them, and such data are naturally represented as graphs. In particular, the detection of a small subset of vertices with anomalous coordinated connectivity is of broad interest, for problems such as detecting strange traffic in a computer network or unknown communities in a social network. These problems become more difficult as the background graph grows larger and noisier and the coordination patterns become more subtle. In this paper, we discuss the computational challenges of a statistical framework designed to address this cross-mission challenge. The statistical framework is based on spectral analysis of the graph data, and three partitioning methods are evaluated for computing the principal eigenvector of the graph's residuals matrix. While a standard one-dimensional partitioning technique enables this computation for up to four billion vertices, the communication overhead prevents this method from being used for even larger graphs. Recent two-dimensional partitioning methods are shown to have much more favorable scaling properties. A data-dependent partitioning method, which has the best scaling performance, is also shown to improve computation time even as a graph changes over time, allowing amortization of the upfront cost.
READ LESS

Summary

Numerous applications focus on the analysis of entities and the connections between them, and such data are naturally represented as graphs. In particular, the detection of a small subset of vertices with anomalous coordinated connectivity is of broad interest, for problems such as detecting strange traffic in a computer network...

READ MORE
Sparse matrix partitioning for parallel eigenanalysis of large static and dynamic graphs

Genetic sequence matching using D4M big data approaches

September 9, 2014
  |
Conference Paper
Author:
Darrell O. Ricke
…
Published in:
HPEC 2014: IEEE Conf. on High Performance Extreme Computing, 9-11 September 2014.
Topic:
high performance computing
R&D area:
R&D group:

Summary

Recent technological advances in Next Generation Sequencing tools have led to increasing speeds of DNA sample collection, preparation, and sequencing. One instrument can produce over 600 Gb of genetic sequence data in a single run. This creates new opportunities to efficiently handle the increasing workload. We propose a new method of fast genetic sequence analysis using the Dynamic Distributed Dimensional Data Model (D4M) - an associative array environment for MATLAB developed at MIT Lincoln Laboratory. Based on mathematical and statistical properties, the method leverages big data techniques and the implementation of an Apache Acculumo database to accelerate computations one-hundred fold over other methods. Comparisons of the D4M method with the current gold-standard for sequence analysis, BLAST, show the two are comparable in the alignments they find. This paper will present an overview of the D4M genetic sequence algorithm and statistical comparisons with BLAST.
READ LESS

Summary

Recent technological advances in Next Generation Sequencing tools have led to increasing speeds of DNA sample collection, preparation, and sequencing. One instrument can produce over 600 Gb of genetic sequence data in a single run. This creates new opportunities to efficiently handle the increasing workload. We propose a new method...

READ MORE
Genetic sequence matching using D4M big data approaches

Achieving 100,000,000 database inserts per second using Accumulo and D4M

September 9, 2014
  |
Conference Paper
Author:
Jeremy Kepner
…
Published in:
HPEC 2014: IEEE Conf. on High Performance Extreme Computing, 9-11 September 2014.
Topic:
big data
R&D area:
R&D group:

Summary

The Apache Accumulo database is an open source relaxed consistency database that is widely used for government applications. Accumulo is designed to deliver high performance on unstructured data such as graphs of network data. This paper tests the performance of Accumulo using data from the Graph500 benchmark. The Dynamic Distributed Dimensional Data Model (D4M) software is used to implement the benchmark on a 216-node cluster running the MIT SuperCloud software stack. A peak performance of over 100,000,000 database inserts per second was achieved which is 100x larger than the highest previously published value for any other database. The performance scales linearly with the number of ingest clients, number of database servers, and data size. The performance was achieved by adapting several supercomputing techniques to this application: distributed arrays, domain decomposition, adaptive load balancing, and single-program-multiple-data programming.
READ LESS

Summary

The Apache Accumulo database is an open source relaxed consistency database that is widely used for government applications. Accumulo is designed to deliver high performance on unstructured data such as graphs of network data. This paper tests the performance of Accumulo using data from the Graph500 benchmark. The Dynamic Distributed...

READ MORE
Achieving 100,000,000 database inserts per second using Accumulo and D4M
641-650 of 2318