Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “SPEECH”

Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 19 records
All sourcesNASA NTRSDOE OSTI
DOE OSTI2014

Patient-Reported Voice and Speech Outcomes After Whole-Neck Intensity Modulated Radiation Therapy and Chemotherapy for Oropharyngeal Cancer: Prospective Longitudinal Study

Purpose: To describe voice and speech quality changes and their predictors in patients with locally advanced oropharyngeal cancer treated on prospective clinical studies of organ-preserving chemotherapy–intensity modulated radiation therapy (chemo-IMRT). Methods and Materials: Ninety-one patients with stage III/IV oropharyngeal cancer were treated on 2 consecutive prospective studies of definitive chemoradiation using whole-field IMRT from 2003 to 2011. Patient-reported voice and speech quality were longitudinally assessed from before treatment through 24 months using the Communication Domain of the Head and Neck Quality of Life (HNQOL-C) instrument and the Speech question of the University of Washington Quality of Life (UWQOL-S) instrument, respectively. Factors associated with patient-reported voice quality worsening from baseline and speech impairment were assessed. Results: Voice quality decreased maximally at 1 month, with 68% and 41% of patients reporting worse HNQOL-C and UWQOL-S scores compared with before treatment, and improved thereafter, recovering to baseline by 12-18 months on average. In contrast, observer-rated larynx toxicity was rare (7% at 3 months; 5% at 6 months). Among patients with mean glottic larynx (GL) dose ≤20 Gy, >20-30 Gy, >30-40 Gy, >40-50 Gy, and >50 Gy, 10%, 32%, 25%, 30%, and 63%, respectively, reported worse voice quality at 12 months compared with before treatment (P=.011). Results for speech impairment were similar. Glottic larynx dose, N stage, neck dissection, oral cavity dose, and time since chemo-IMRT were univariately associated with either voice worsening or speech impairment. On multivariate analysis, mean GL dose remained independently predictive for both voice quality worsening (8.1%/Gy) and speech impairment (4.3%/Gy). Conclusions: Voice quality worsening and speech impairment after chemo-IMRT for locally advanced oropharyngeal cancer were frequently reported by patients, underrecognized by clinicians, and independently associated with GL dose. These findings support reducing mean GL dose to as low as reasonably achievable, aiming at ≤20 Gy when the larynx is not a target.

62 RADIOLOGY AND NUCLEAR MEDICINE↗
DOE OSTI2016

Neural speech recognition: continuous phoneme decoding using spatiotemporal representations of human cortical activity

The superior temporal gyrus (STG) and neighboring brain regions play a key role in human language processing. Previous studies have attempted to reconstruct speech information from brain activity in the STG, but few of them incorporate the probabilistic framework and engineering methodology used in modern speech recognition systems. Here, in this work, we describe the initial efforts toward the design of a neural speech recognition (NSR) system that performs continuous phoneme recognition on English stimuli with arbitrary vocabulary sizes using the high gamma band power of local field potentials in the STG and neighboring cortical areas obtained via electrocorticography. The system implements a Viterbi decoder that incorporates phoneme likelihood estimates from a linear discriminant analysis model and transition probabilities from an n-gram phonemic language model. Grid searches were used in an attempt to determine optimal parameterizations of the feature vectors and Viterbi decoder. The performance of the system was significantly improved by using spatiotemporal representations of the neural activity (as opposed to purely spatial representations) and by including language modeling and Viterbi decoding in the NSR system. These results emphasize the importance of modeling the temporal dynamics of neural responses when analyzing their variations with respect to varying stimuli and demonstrate that speech recognition techniques can be successfully leveraged when decoding speech from neural signals. Guided by the results detailed in this work, further development of the NSR system could have applications in the fields of automatic speech recognition and neural prosthetics.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2022

Speaker-targeted Synthetic Speech Detection

Text-to-speech technologies are evolving quickly towards realistic-sounding human-like voices. As this technology improves, so does the opportunity for malpractice in speaker identification (SID) via spoofing, the process of impersonating a voice biometric via synthesis. More data typically equates to a more realistic voice model, which poses an issue for well-known subjects, such as politicians and celebrities, who have vast amounts of multimedia available online. Detection of synthetic speech has relied on signal processing techniques that focus on the generation of new acoustic features and train deep learning models to detect when an audio file has been manipulated through the characterization of unnatural changes or artifacts. However, these techniques do not use any information from the speaker they are evaluating. This paper proposes to incorporate information from the speaker-of-interest (SoI) into the models to avoid specific spoofing attacks for certain vulnerable people. The wealth of data for well-known people can also be used to train a speaker-specific spoofing detector with a higher level of accuracy than a speaker-independent model. The paper proposes a new xResNet-PLDA system and compares it to three different baseline systems: a state-of-the-art speaker identification system, an xResNet system trained to discriminate between bona fide and fake speech, and a speaker identification system in which the PLDA and calibration models were trained with bona fide and fake speech. We evaluated the systems in two different scenarios — a cross-validation scenario and a hold-out scenario — with three different databases. We show how the proposed system outperforms dramatically the baseline systems in each scenario and for each database. Finally, we show how using a small amount of the SoI’s speech to adapt global calibration parameters improves the performance of the system, especially in unseen conditions.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗
DOE OSTI2019

Deep learning as a tool for neural data analysis: Speech classification and cross-frequency coupling in human sensorimotor cortex

A fundamental challenge in neuroscience is to understand what structure in the world is represented in spatially distributed patterns of neural activity from multiple single-trial measurements. This is often accomplished by learning a simple, linear transformations between neural features and features of the sensory stimuli or motor task. While successful in some early sensory processing areas, linear mappings are unlikely to be ideal tools for elucidating nonlinear, hierarchical representations of higher-order brain areas during complex tasks, such as the production of speech by humans. Here, we apply deep networks to predict produced speech syllables from a dataset of high gamma cortical surface electric potentials recorded from human sensorimotor cortex. We find that deep networks had higher decoding prediction accuracy compared to baseline models. Having established that deep networks extract more task relevant information from neural data sets relative to linear models (i.e., higher predictive accuracy), we next sought to demonstrate their utility as a data analysis tool for neuroscience. We first show that deep network's confusions revealed hierarchical latent structure in the neural data, which recapitulated the underlying articulatory nature of speech motor control. We next broadened the frequency features beyond high-gamma and identified a novel high-gamma-to-beta coupling during speech production. Finally, we used deep networks to compare task-relevant information in different neural frequency bands, and found that the high-gamma band contains the vast majority of information relevant for the speech prediction task, with little-to-no additional contribution from lower-frequency amplitudes. Together, these results demonstrate the utility of deep networks as a data analysis tool for basic and applied neuroscience.

60 APPLIED LIFE SCIENCES↗
DOE OSTI2018

Retrospective Analysis of Clinical Performance of an Estonian Speech Recognition System for Radiology: Effects of Different Acoustic and Language Models

The aim of this study was to analyze retrospectively the influence of different acoustic and language models in order to determine the most important effects to the clinical performance of an Estonian language-based non-commercial radiology-oriented automatic speech recognition (ASR) system. An ASR system was developed for Estonian language in radiology domain by utilizing open-source software components (Kaldi toolkit, Thrax). The ASR system was trained with the real radiology text reports and dictations collected during development phases. The final version of the ASR system was tested by 11 radiologists who dictated 219 reports in total, in spontaneous manner in a real clinical environment. The audio files collected in the final phase were used to measure the performance of different versions of the ASR system retrospectively. ASR system versions were evaluated by word error rate (WER) for each speaker and modality and by WER difference for the first and the last version of the ASR system. Total average WER for the final version throughout all material was improved from 18.4% of the first version (v1) to 5.8% of the last (v8) version which corresponds to relative improvement of 68.5%. WER improvement was strongly related to modality and radiologist. In summary, the performance of the final ASR system version was close to optimal, delivering similar results to all modalities and being independent on user, the complexity of the radiology reports, user experience, and speech characteristics.

62 RADIOLOGY AND NUCLEAR MEDICINE↗
DOE OSTI2025

Automatic speech recognition predicts contemporaneous earthquake fault displacement

Abstract Significant progress has been made in probing the state of an earthquake fault by applying machine learning to continuous seismic waveforms. The breakthroughs were originally obtained from laboratory shear experiments and numerical simulations of fault shear, then successfully extended to slow-slipping faults. Here we apply the Wav2Vec-2.0 self-supervised framework for automatic speech recognition to continuous seismic signals emanating from a sequence of moderate magnitude earthquakes during the 2018 caldera collapse at the Kīlauea volcano on the island of Hawai’i. We pre-train the Wav2Vec-2.0 model using caldera seismic waveforms and augment the model architecture to predict contemporaneous surface displacement during the caldera collapse sequence, a proxy for fault displacement. We find the model displacement predictions to be excellent. The model is adapted for near-future prediction information and found hints of prediction capability, but the results are not robust. The results demonstrate that earthquake faults emit seismic signatures in a similar manner to laboratory and numerical simulation faults, and artificial intelligence models developed for encoding audio of speech may have important applications in studying active fault zones.

58 GEOSCIENCES↗
DOE OSTI2023

Measuring Equality in Machine Learning Security Defenses: A Case Study in Speech Recognition

Over the past decade, the machine learning security community has developed a myriad of defenses for evasion attacks. An understudied question in that community is: for whom do these defenses defend? This work considers common approaches to defending learned systems and how security defenses result in performance inequities across different sub-populations. We outline appropriate parity metrics for analysis and begin to answer this question through empirical results of the fairness implications of machine learning security methods. We find that many methods that have been proposed can cause direct harm, like false rejection and unequal benefits from robustness training. The framework we propose for measuring defense equality can be applied to robustly trained models, preprocessing-based defenses, and rejection methods. We identify a set of datasets with a user-centered application and a reasonable computational cost suitable for case studies in measuring the equality of defenses. In our case study of speech command recognition, we show how such adversarial training and augmentation have non-equal but complex protections for social subgroups across gender, accent, and age in relation to user coverage. We present a comparison of equality between two rejection-based defenses: randomized smoothing and neural rejection, finding randomized smoothing more equitable due to the sampling mechanism for minority groups. This represents the first work examining the disparity in the adversarial robustness in the speech domain and the fairness evaluation of rejection-based defenses.

• Artificial intelligence (AI) / machine learning ↗
DOE OSTI2016

Constrained Versions of DEDICOM for Use in Unsupervised Part-Of-Speech Tagging

This reports describes extensions of DEDICOM (DEcomposition into DIrectional COMponents) data models [3] that incorporate bound and linear constraints. The main purpose of these extensions is to investigate the use of improved data models for unsupervised part-of-speech tagging, as described by Chew et al. [2]. In that work, a single domain, two-way DEDICOM model was computed on a matrix of bigram fre- quencies of tokens in a corpus and used to identify parts-of-speech as an unsupervised approach to that problem. An open problem identi ed in that work was the com- putation of a DEDICOM model that more closely resembled the matrices used in a Hidden Markov Model (HMM), speci cally through post-processing of the DEDICOM factor matrices. The work reported here consists of the description of several models that aim to provide a direct solution to that problem and a way to t those models. The approach taken here is to incorporate the model requirements as bound and lin- ear constrains into the DEDICOM model directly and solve the data tting problem as a constrained optimization problem. This is in contrast to the typical approaches in the literature, where the DEDICOM model is t using unconstrained optimization approaches, and model requirements are satis ed as a post-processing step.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2018

Seismic Phase Identification with Speech Recognition Algorithms

Seismic signals are composed of the seismic waves (phases) that reach a sensor, similar to the way speech signals are composed of phonemes that reach a listener's ear. Large/small seismic events near/far from a sensor are similar to loud/quiet speakers with high/low-pitched voices. We leverage ideas from speech recognition for the classification of seismic phases at a seismic sensor. Seismic Phase ID is challenging due to the varying paths and distances an event takes to reach a sensor, but there is consistent structure of the makeup (e.g. ordering) of the different phases arriving at the sensor.

58 GEOSCIENCES↗
DOE OSTI2016

2015 Nuclear Fusion Prize acceptance speech

This is the 2015 Nuclear Fusion Prize acceptance speech of R.J. Goldston: It is a great pleasure to receive the 2015 Nuclear Fusion award for my work developing a heuristic drift-based model for the power scrape-off width in tokamaks. I was particularly pleased to receive the award from IAEA Director General Yukiya Amano, whose thoughtful leadership has advanced the cause of nuclear non-proliferation mightily.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗
DOE OSTI2019

On finding balance between collaborative noise and speech privacy in open offices

A growing number of businesses are moving towards open-plan offices as a way of encouraging impromptu collaborative problem solving among workers. However, while collaboration increases in open-plan offices, a commensurate increase in general noise can hinder employees that need quieter conditions to prosper. In this paper the effect of conversational noise within an open plan environment are quantified, and the degree to which noise restrictions compromise problem-solving examined. Measurements indicate that the noise from six people talking in a 1,000 sq ft room was sufficient to meet recommended levels under ASHRAE guidelines. This suggests an appropriate density of spontaneous collaborative talk in open offices. Subject surveys also indicated an increase in non-verbal collaboration in later tests, suggesting that with thought and careful design a balance between collaboration and quiet can be obtained.

open-plan, collaboration, office design, speech pr↗
DOE OSTI2020

Watch former Lab Director and Manhattan Project Veteran Harold Agnew’s 1985 speech

From the first nuclear reactor to the weapon development to the bomb delivery in war, Harold Agnew was there. The only two atomic bombs to ever be used in combat – Little Boy, a uranium gun-type weapon, and Fat Man, a plutonium implosion weapon – were released above Japan 75 years ago on Aug. 6 and Aug. 9, respectively. World War II ended not long after. And, according to The New York Times, Agnew was the only person to witness the whole undertaking, from reactor to Hiroshima: Years earlier, on a frigid December day in 1942 in Chicago, a 21-year-old Agnew was one of a few dozen gathered to see atoms split in two. Not long after, Agnew came to Los Alamos as a graduate student to work on the development the atomic bomb. Then, on Aug. 6, 1945, Agnew was a part of the delivery of Little Boy; he flew as a scientific member on the mission to Hiroshima to film the mushroom cloud and measure the shockwave. It is the only existing footage of the Hiroshima bombing from the air; other footage was taken, but was most likely not handled properly and was unusable, according to Agnew.

99 GENERAL AND MISCELLANEOUS↗