Publications
A unified deep neural network for speaker and language recognition
September 6, 2015
Conference Paper
Published in:
INTERSPEECH 2015: 15th Annual Conf. of the Int. Speech Communication Assoc., 6-10 September 2015.
Topic:
R&D area:
Summary
Significant performance gains have been reported separately for speaker recognition (SR) and language recognition (LR) tasks using either DNN posteriors of sub-phonetic units or DNN feature representations, but the two techniques have not been compared on the same SR or LR task or across SR and LR tasks using the same DNN. In this work we present the application of a single DNN for both tasks using the 2013 Domain Adaptation Challenge speaker recognition (DAC13) and the NIST 2011 language recognition evaluation (LRE11) benchmarks. Using a single DNN trained on Switchboard data we demonstrate large gains in performance on both benchmarks: a 55% reduction in EER for the DAC13 out-of-domain condition and a 48% reduction in Cavg on the LRE11 30s test condition. Score fusion and feature fusion are also investigated as is the performance of the DNN technologies at short durations for SR.
Summary
Significant performance gains have been reported separately for speaker recognition (SR) and language recognition (LR) tasks using either DNN posteriors of sub-phonetic units or DNN feature representations, but the two techniques have not been compared on the same SR or LR task or across SR and LR tasks using the...
READ MORE
Estimating lower vocal tract features with closed-open phase spectral analyses
September 6, 2015
Conference Paper
Published in:
INTERSPEECH 2015: 15th Annual Conf. of the Int. Speech Communication Assoc., 6-10 September 2015.
Summary
Previous studies have shown that, in addition to being speaker-dependent yet context-independent, lower vocal tract acoustics significantly impact the speech spectrum at mid-to-high frequencies (e.g 3-6kHz). The present work automatically estimates spectral features that exhibit acoustic properties of the lower vocal tract. Specifically aiming to capture the cyclicity property of the epilarynx tube, a novel multi-resolution approach to spectral analyses is presented that exploits significant differences between the closed and open phases of a glottal cycle. A prominent null linked to the piriform fossa is also estimated. Examples of the feature estimation on natural speech of the VOICES multi-speaker corpus illustrate that a salient spectral pattern indeed emerges between 3-6kHz across all speakers. Moreover, the observed pattern is consistent with that canonically shown for the lower vocal tract in previous works. Additionally, an instance of a speaker's formant (i.e. spectral peak around 3kHz that has been well-established as a characteristic of voice projection) is quantified here for the VOICES template speaker in relation to epilarynx acoustics. The corresponding peak is shown to be double the power on average compared to the other speakers (20 vs 10 dB).
Summary
Previous studies have shown that, in addition to being speaker-dependent yet context-independent, lower vocal tract acoustics significantly impact the speech spectrum at mid-to-high frequencies (e.g 3-6kHz). The present work automatically estimates spectral features that exhibit acoustic properties of the lower vocal tract. Specifically aiming to capture the cyclicity property of...
READ MORE
A near-quantum-limited Josephson traveling-wave parametric amplifier
September 3, 2015
Journal Article
Published in:
Sci., Vol. 350, No. 6258, 16 October 2015,pp. 307-10.
Summary
Detecting single photon level signals--carriers of both classical and quantum information--is particularly challenging for low-energy microwave frequency excitations. Here we introduce a superconducting amplifier based on a Josephson junction transmission line. Unlike current standing-wave parametric amplifiers, this traveling wave architecture robustly achieves high gain over a bandwidth of several gigahertz with sufficient dynamic range to read out 20 superconducting qubits. To achieve this performance, we introduce a sub-wavelength resonant phase matching technique that enables the creation of nonlinear microwave devices with unique dispersion relations. We benchmark the amplifier with weak measurements, obtaining a high quantum efficiency of 75% (70% including following amplifier noise). With a flexible design based on compact lumped elements, this Josephson amplifier has broad applicability to microwave metrology and quantum optics.
Summary
Detecting single photon level signals--carriers of both classical and quantum information--is particularly challenging for low-energy microwave frequency excitations. Here we introduce a superconducting amplifier based on a Josephson junction transmission line. Unlike current standing-wave parametric amplifiers, this traveling wave architecture robustly achieves high gain over a bandwidth of several gigahertz...
READ MORE
Reduction of trapped-ion anomalous heating by in situ surface plasma cleaning
August 28, 2015
Journal Article
Published in:
Phys. Rev. A, At. Mol. Opt. Phys., Vol. 92, No. 2, 2015, 020302.
Summary
Anomalous motional heating is a major obstacle to scalable quantum information processing with trapped ions. Although the source of this heating is not yet understood, several previous studies suggest that noise due to surface contaminants is the limiting heating mechanism in some instances. We demonstrate an improvement by a factor of 4 in the room-temperature heating rate of a niobium surface electrode trap by in situ plasma cleaning of the trap surface. This surface treatment was performed with a simple homebuilt coil assembly and commercially available matching network and is considerably gentler than other treatments, such as ion milling or laser cleaning, that have previously been shown to improve ion heating rates. We do not see an improvement in the heating rate when the trap is operated at cryogenic temperatures, pointing to a role of thermally activated surface contaminants in motional heating whose activity may freeze out at low temperatures.
Summary
Anomalous motional heating is a major obstacle to scalable quantum information processing with trapped ions. Although the source of this heating is not yet understood, several previous studies suggest that noise due to surface contaminants is the limiting heating mechanism in some instances. We demonstrate an improvement by a factor...
READ MORE
A spectral framework for anomalous subgraph detection
August 15, 2015
Journal Article
Published in:
IEEE Trans. Signal Process., Vol. 63, No. 16, 15 August 2015, 4191-4206.
Topic:
R&D area:
Summary
A wide variety of application domains is concerned with data consisting of entities and their relationships or connections, formally represented as graphs. Within these diverse application areas, a common problem of interest is the detection of a subset of entities whose connectivity is anomalous with respect to the rest of the data. While the detection of such anomalous subgraphs has received a substantial amount of attention, no application-agnostic framework exists for analysis of signal detectability in graph-based data. In this paper, we describe a framework that enables such analysis using the principal eigenspace of a graph's residuals matrix, commonly called the modularity matrix in community detection. Leveraging this analytical tool, we show that the framework has a natural power metric in the spectral norm of the anomalous subgraph's adjacency matrix (signal power) and of the background graph's residuals matrix (noise power). We propose several algorithms based on spectral properties of the residuals matrix, with more computationally expensive techniques providing greater detection power. Detection and identification performance are presented for a number of signal and noise models, including clusters and bipartite foregrounds embedded into simple random backgrounds, as well as graphs with community structure and realistic degree distributions. The trends observed verify intuition gleaned from other signal processing areas, such as greater detection power when the signal is embedded within a less active portion of the background. We demonstrate the utility of the proposed techniques in detecting small, highly anomalous subgraphs in real graphs derived from Internet traffic and product co-purchases.
Summary
A wide variety of application domains is concerned with data consisting of entities and their relationships or connections, formally represented as graphs. Within these diverse application areas, a common problem of interest is the detection of a subset of entities whose connectivity is anomalous with respect to the rest of...
READ MORE
Super-resolution microscopy by movable thin-films with embedded microspheres: resolution analysis
August 14, 2015
Journal Article
Published in:
Ann. Phys., Vol. 527, No. 7-8, 2015, pp. 513-22.
Summary
Microsphere-assisted imaging has emerged as an extraordinary simple technique of obtaining optical super-resolution. This work addresses two central problems in developing this technology: i) methodology of the resolution measurements and ii) limited field-of-view provided by each sphere. It is suggested that a standard method of resolution analysis in far-field microscopy based on convolution with the point-spread function can be extended into the superresolution area. This allows developing a unified approach to resolution measurements, which can be used for comparing results obtained by different techniques. To develop the surface scanning functionality, the high-index (n ~ 2) barium titanate glass microspheres were embedded in polydimethylsiloxane (PDMS) thin-films. It is shown that such films adhere to the surface of nanoplasmonic structures so that the tips of embedded spheres experience the objects' optical near-fields. Based on rigorous criteria, the resolution ~lambda/6-lambda/7 (where lambda is the illumination wavelength) is demonstrated for arrays of Au dimers and bowties. Such films can be translated along the surface of investigated samples after liquid lubrication. It is shown that just after lubrication the resolution is diffraction limited, however the super-resolution gradually recovers as the lubricant evaporates.
Summary
Microsphere-assisted imaging has emerged as an extraordinary simple technique of obtaining optical super-resolution. This work addresses two central problems in developing this technology: i) methodology of the resolution measurements and ii) limited field-of-view provided by each sphere. It is suggested that a standard method of resolution analysis in far-field microscopy...
READ MORE
Iris biometric security challenges and possible solutions: for your eyes only? Using the iris as a key
August 13, 2015
Journal Article
Published in:
IEEE Signal Process. Mag., Vol. 32, No. 5, September 2015, pp. 42-53.
Summary
Biometrics were originally developed for identification, such as for criminal investigations. More recently, biometrics have been also utilized for authentication. Most biometric authentication systems today match a user's biometric reading against a stored reference template generated during enrollment. If the reading and the template are sufficiently close, the authentication is considered successful and the user is authorized to access protected resources. This binary matching approach has major inherent vulnerabilities. An alternative approach to biometric authentication proposes to use fuzzy extractors (also known as biometric cryptosystems), which derive cryptographic keys from noisy sources, such as biometrics. In theory, this approach is much more robust and can enable cryptographic authorization. Unfortunately, for many biometrics that provide high-quality identification, fuzzy extractors provide no security guarantees. This gap arises in part because of an objective mismatch. The quality of a biometric identification is typically measured using false match rate (FMR) versus false nonmatch rate (FNMR). As a result, biometrics have been extensively optimized for this metric. However, this metric says little about the suitability of a biometric for key derivation. In this article, we illustrate a metric that can be used to optimize biometrics for authentication. Using iris biometrics as an example, we explore possible directions for improving processing and representation according to this metric. Finally, we discuss why strong biometric authentication remains a challenging problem and propose some possible future directions for addressing these challenges.
Summary
Biometrics were originally developed for identification, such as for criminal investigations. More recently, biometrics have been also utilized for authentication. Most biometric authentication systems today match a user's biometric reading against a stored reference template generated during enrollment. If the reading and the template are sufficiently close, the authentication is...
READ MORE
Operational exercise integration recommendations for DoD cyber ranges
August 5, 2015
Technical Report
Published in:
MIT Lincoln Laboratory Report TR-1187
Topic:
R&D area:
R&D group:
Summary
Cyber-enabled and cyber-physical systems connect and engage virtually every mission-critical military capability today. And as more warfighting technologies become integrated and connected, both the risks and opportunities from a cyberwarfare continue to grow--motivating sweeping requirements and investments in cybersecurity assessment capabilities to evaluate technology vulnerabilities, operational impacts, and operator effectiveness. Operational testing of cyber capabilities, often in conjunction with major military exercises, provides valuable connections to and feedback from the operational warfighter community. These connections can help validate capability impact on the mission and, when necessary, provide course-correcting feedback to the technology development process and its stakeholders. However, these tests are often constrained in scope, duration, and resources and require a thorough and holistic approach, especially with respect to cyber technology assessments, where additional safety and security constraints are often levied. This report presents a summary of the state of the art in cyber assessment technologies and methodologies and prescribes an approach to the employment of cyber range operational exercises (OPEXs). Numerous recommendations on general cyber assessment methodologies and cyber range design are included, the most significant of which are summarized below. -Perform bottom-up and top-down assessment formulation methodologies to robustly link mission and assessment objectives to metrics, success criteria, and system observables. -Include threat-based assessment formulation methodologies that define risk and security metrics within the context of mission-relevant adversarial threats and mission-critical system assets. -Follow a set of cyber range design mantras to guide and grade the design of cyber range components. -Call for future work in live-to-virtual exercise integration and cross-domain modeling and simulation technologies. - Call for continued integration of developmental and operational cyber assessment events, development of reusable cyber assessment test tools and processes, and integration of a threat-based assessment approach across the cyber technology acquisition cycle. Finally, this recommendations report was driven by observations made by the MIT Lincoln Laboratory (MIT LL) Cyber Measurement Campaign (CMC) team during an operational demonstration event for the DoD Enterprise Cyber Range Environment (DECRE) Command and Control Information Systems (C2IS). This report also incorporates a prior CMC report based on Pacific Command (PACOM) exercise observations, as well as MIT LL's expertise in cyber range development and cyber systems assessment.
Summary
Cyber-enabled and cyber-physical systems connect and engage virtually every mission-critical military capability today. And as more warfighting technologies become integrated and connected, both the risks and opportunities from a cyberwarfare continue to grow--motivating sweeping requirements and investments in cybersecurity assessment capabilities to evaluate technology vulnerabilities, operational impacts, and operator effectiveness...
READ MORE
Comparison of gate dielectric plasma damage from plasma-enhanced atomic layer deposited and magnetron sputtered TiN metal gates
July 29, 2015
Journal Article
Published in:
J. Appl. Phys., Vol. 118, No. 4, 2015, 045307.
Summary
Fully depleted silicon-on-insulator transistors were fabricated using two different metal gate deposition mechanisms to compare plasma damage effects on gate oxide quality. Devices fabricated with both plasma-enhanced atomic-layer-deposited (PE-ALD) TiN gates and magnetron plasma sputtered TiN gates showed very good electrostatics and short-channel characteristics. However, the gate oxide quality was markedly better for PE-ALD TiN. A significant reduction in interface state density was inferred from capacitance-voltage measurements as well as a 1200 x reduction in gate leakage current. A high-power magnetron plasma source produces a much higher energetic ion and vacuum ultra-violet (VUV) photon flux to the wafer compared to a low-power inductively coupled PE-ALD source. The ion and VUV photons produce defect states in the bulk of the gate oxide as well as at the oxide-silicon interface, causing higher leakage and potential reliability degradation.
Summary
Fully depleted silicon-on-insulator transistors were fabricated using two different metal gate deposition mechanisms to compare plasma damage effects on gate oxide quality. Devices fabricated with both plasma-enhanced atomic-layer-deposited (PE-ALD) TiN gates and magnetron plasma sputtered TiN gates showed very good electrostatics and short-channel characteristics. However, the gate oxide quality was...
READ MORE
Versatile alignment layer method for new types of liquid crystal photonic devices
July 21, 2015
Journal Article
Published in:
J. Appl. Phys., Vol. 118, No. 3, July 2015, pp. 034501.
Topic:
R&D area:
Summary
Liquid crystal photonic devices are becoming increasingly popular. These devices often present a challenge when it comes to creating a robust alignment layer in pre-assembled cells. In this paper, we describe a method of infusing a dye into a microcavity to produce an effective photo-definable alignment layer. However, previous research on such alignment layers has shown that they have limited stability, particularly against subsequent light exposure. As such, we further describe a method of utilizing a pre-polymer, infused into the microcavity along with the liquid crystal, to provide photostability. We demonstrate that the polymer layer, formed under ultraviolet irradiation of liquid crystal cells, has been effectively localized to a thin region near the substrate surface and provides a significant improvement in the photostability of the liquid crystal alignment. This versatile alignment layer method, capable of being utilized in devices from the described microcavities to displays, offers significant promise for new photonics applications.
Summary
Liquid crystal photonic devices are becoming increasingly popular. These devices often present a challenge when it comes to creating a robust alignment layer in pre-assembled cells. In this paper, we describe a method of infusing a dye into a microcavity to produce an effective photo-definable alignment layer. However, previous research...
READ MORE