Search

The type of voice model used in Computer Assisted Pronunciation Instruction is a crucial factor in the quality of practice and the amount of uptake by language learners. As an example, prior research indicates that second-language learners are more likely to succeed when they imitate a speaker with a voice similar to their own, a so-called “golden speaker”. This manuscript presents Golden Speaker Builder (GSB), a tool that allows learners to generate a personalized “golden-speaker” voice: one that mirrors their own voice but with a native accent. We describe the overall system design, including the web application with its user interface, and the underlying speech analysis/synthesis algorithms. Next, we present results from a series of listening tests, which show that GSB is capable of synthesizing such golden-speaker voices. Finally, we present results from a user study in a language-instruction setting, which show that practising with GSB leads to improved fluency and comprehensibility. We suggest reasons for why learners improved as they did and recommendations for the next iteration of the training.

Diet monitoring is an essential intervention component for a number of diseases, from type 2 diabetes to cardiovascular diseases. However, current methods for diet monitoring are burdensome and often inaccurate. In prior work, we showed that continuous glucose monitors (CGMs) may be used to predict meal macronutrients (e.g., carbohydrates, protein, fat) by analyzing the shape of the post-prandial glucose response. In this study, we examine a number of additional dietary biomarkers in blood by their ability to improve macronutrient prediction, compared to using CGMs alone. For this purpose, we conducted a nutritional study where (n = 10) participants consumed nine different mixed meals with varied but known macronutrient amounts, and we analyzed the concentration of 33 dietary biomarkers (including amino acids, insulin, triglycerides, and glucose) at various times post-prandially. Then, we built machine learning models to predict macronutrient amounts from (1) individual biomarkers and (2) their combinations. We find that the additional blood biomarkers provide complementary information, and more importantly, achieve lower normalized root mean squared error (NRMSE) for the three macronutrients (carbohydrates: 22.9%; protein: 23.4%; fat: 32.3%) than CGMs alone (carbohydrates: 28.9%, t(18) =1.64, p =0.060; protein: 46.4%, t(18) =5.38, p 0.001; fat: 40.0%, t(18) =2.09, p =0.025). Our main conclusion is that augmenting CGMs to measure these additional dietary biomarkers improves macronutrient prediction performance, and may ultimately lead to the development of automated methods to monitor nutritional intake. This work is significant to biomedical research as it provides a potential solution to the long-standing problem of diet monitoring, facilitating new interventions for a number of diseases.

Mid-infrared (mid-IR) sensors consisting of silicon nitride (SiN) waveguides were designed and tested to detect volatile organic compounds (VOCs). SiN thin films, prepared by low-pressure chemical vapor deposition (LPCVD), have a broad mid-IR transparent region and a lower refractive index (nSiN = 2.0) than conventional materials such as Si (nSi = 3.4), which leads to a stronger evanescent wave and therefore higher sensitivity, as confirmed by a finite-difference eigenmode (FDE) calculation. Further, in-situ monitoring of three VOCs (acetone, ethanol, and isoprene) was experimentally demonstrated through characteristic absorption measurements at wavelengths λ = 3.0–3.6 μm. The SiN waveguide showed a five-fold sensitivity improvement over the Si waveguide due to its stronger evanescent field. To our knowledge, this is the first time SiN waveguides are used to perform on-chip mid-IR spectral measurements for VOC detection. Thus, the developed waveguide sensor has the potential to be used as a compact device module capable of monitoring multiple gaseous analytes for health, agricultural and environmental applications.

Managing diabetes requires following a healthy lifestyle, including monitoring dietary intake. Prior work has shown that meals with different macronutrient composition can have distinct postprandial glucose responses (PPGR), therefore suggesting that PPGRs may be used to monitor diet automatically. Yet, PPGRs shown large variability across individuals. This paper proposes a metric-learning approach to achieve personalized meal macronutrient estimation from PPGRs. The metric learning approach utilizes a Siamese neural network (SNN) architecture, which learns a PPGR embedding via a contrastive loss function adapted to the task of interest. Specifically, the proposed contrastive loss is designed so that it maximizes the distance between meals of similar macronutrient composition and minimizes the distance between meals with different macronutrients. This loss is further computed within each individual, therefore reducing individual differences in PPGRs. Our results show that the proposed metric learning approach outperforms a feedforward neural network when estimating the amount of protein, carbohydrate, and fat in a meal. These suggest the feasibility of using PPGRs to track meal macronutrient composition, supporting dietary informatics applications for precision health and nutrition.

Measuring dietary intake is a major challenge in the management of chronic diseases. Current methods rely on self-report measures, which are cumbersome to obtain and often unreliable. This article presents an approach to estimate dietary intake automatically by analyzing the post-prandial glucose response (PPGR) of a meal, as measured with continuous glucose monitors. In particular, we propose a sparse-coding technique that can be used to estimate the amounts of macronutrients (carbohydrates, protein, fat) in a meal from the meal’s PPGR. We use Lasso regularization to represent the PPGR of a new meal as a sparse combination of PPGRs in a dictionary, then combine the sparse weights with the macronutrient amounts in the dictionary’s meals to estimate the macronutrients in the new meal. We evaluate the approach on a dataset containing nine standardized meals and their corresponding PPGRs, consumed by fifteen participants. The proposed technique consistently outperforms two baseline systems based on ridge regression and nearest-neighbors, in terms of correlation and normalized root mean square error of the predictions.

Diet monitoring is an important component of interventions in type 2 diabetes, but is time intensive and often inaccurate. To address this issue, we describe an approach to monitor diet automatically, by analyzing fluctuations in glucose after a meal is consumed. In particular, we evaluate three standardization techniques (baseline correction, feature normalization, and model personalization) that can be used to compensate for the large individual differences that exist in food metabolism. Then, we build machine learning models to predict the amounts of macronutrients in a meal from the associated glucose responses. We evaluate the approach on a dataset containing glucose responses for 15 participants who consumed 9 meals. Three techniques improve the accuracy of the models: subtracting the baseline glucose, performing z-score normalization, and scaling the amount of macronutrients by each individuals’ body mass index.

Accent conversion AC aims to transform non-native utterances to sound as if the speaker had a native accent. This can be achieved by mapping source speech spectra from a native speaker into the acoustic space of the target non-native speaker. In prior work, we proposed an AC approach that matches frames between the two speakers based on their acoustic similarity after compensating for differences in vocal tract length. In this paper, we propose a new approach that matches frames between the two speakers based on their phonetic rather than acoustic similarity. Namely, we map frames from the two speakers into a phonetic posteriorgram using speaker-independent acoustic models trained on native speech. We thoroughly evaluate the approach on a speech corpus containing multiple native and non-native speakers. The proposed algorithm outperforms the prior approach, improving ratings of acoustic quality 22% increase in mean opinion score and native accent 69% preference while retaining the voice quality of the non-native speaker. Furthermore, we show that the approach can be used in the reverse conversion direction, i.e., generating speech with a native speaker's voice quality and a non-native accent. Finally, we show that this approach can be applied to non-parallel training data, achieving the same accent conversion performance.

Video sharing sites have become keepers of de-facto digital libraries of sign language content, being used to store videos including the experiences, knowledge, and opinions of many in the deaf or hard of hearing community. Due to limitations of term-based search over metadata, these videos can be difficult to find, reducing their value to the community. Another result is that community members frequently engage in a push-style delivery of content (e.g., emailing or posting links to videos for others in the sign language community) rather than having access be based on the information needs of community members. In prior work, we have shown the potential to detect sign language content using features derived from the video content rather than relying on metadata. Our prior technique was developed with a focus on accuracy of results and are quite computationally expensive, making it unrealistic to apply them on a corpus the size of YouTube or other large video sharing sites. Here, we describe and examine the performance of optimizations that reduce the cost of face detection and the length of video segments processed. We show that optimizations can reduce the computation time required by 96%, while losing only 1% in F1 score. Further, a keyframe-based approach is examined that removes the need to process continuous video. This approach achieves comparable recall but lower precision than the above techniques. Merging the advantages of the optimizations, we also present a staged classifier, where the keyframe approach is used to reduce the number of non-sign language videos fully processed. An analysis of the staged classifier shows a further reduction in average computation time per video while achieving similar quality of results.

This paper compares the effectiveness of two biofeedback mechanisms to promote acquisition and transfer of deep-breathing skills using a casual videogame. The first biofeedback mechanism, game adaptation, delivers respiratory information by altering an internal parameter of the game; the second, visual biofeedback, displays respiratory information explicitly without altering the game. We conduct a user study that examines visual biofeedback and game adaptation as independent variables with electrodermal activity, heart rate variability, and respiration as dependent variables. In particular, we evaluate these forms of biofeedback by their ability to facilitate acquisition of relaxation skills and promote skill transfer to subsequent stressful tasks. Our results indicate that game adaptation promotes skill acquisition and transfer more effectively than visual biofeedback, but that a combination of the two outperforms either in isolation. Combining visual and game biofeedback also results in faster learning of deep-breathing skills than either channel alone. Our study suggests that the two forms of biofeedback play different roles, with game adaptation being more effective in encouraging deep breathing, and visual channels helping players maintain the target breathing rate.

This work focuses on the development of nanoparticle-based layer-by-layer (LbL) coatings for enhancing the detection sensitivity and selectivity of volatile organic compounds (VOCs) using on-chip mid-infrared (MIR) waveguides (WGs). First, we demonstrate construction of conformal coatings of polymer/mesoporous silica nanoparticles (MSNs) on the surface of Si-based WGs using the LbL technique and evaluate the coating deposition conditions, such as pH and substrate withdrawal speed, on the thickness and homogeneity of the assemblies. We then use the modified WGs to achieve enhanced sensitivity and selectivity of polar organic compounds, such as ethanol, versus non-polar ones, such as methane, in the MIR region. In addition, using density functional theory calculations, we show that such an improvement in sensing performance is achieved due to preferential adsorption of ethanol molecules within MSNs in the vicinity of the WG evanescent field.

Changes in land use and increased intensity of rainfall are factors of greater influence that accelerate the soil erosion process. The application of basin scale sedimentological models integrated to Geographic Information Systems (GIS), is a tool that allows the definition of critical zones, and of this the establishment of measures of control of processes of production and transport of sediments. This study evaluates the water erosion using the Universal Soil Loss Equation (USLE), five scenarios were studied: corresponding to the start of operations of the El Portillo II (1980) dam, and the before and after the two extreme events in the basin (September 1998 and October 2005). The transport of sediments was evaluated by regression, using full annually records (eight years) of flows of two gauging stations in the basin. Overall, variations in erosion rates were observed with changes in the vegetal coverage and statistical homogeneity in the gauging data, which allowed adjust them to a regression model, with correlation coefficient upper to 88%.

Water is an essential element in the life of the human being for which its conservation and good use is necessary, in this work we make a proposal for the development of a low-cost system for water management in order to save on Its consumption, using infrared technology, the results provide a mechanism that can be replicated and at different scales, by the use of commonly used devices.

Foreign accent conversion (FAC) aims to create a new voice that has the voice identity of a given second-language (L2) speaker but with a native (L1) accent. Previous FAC approaches usually require training a separate model for each L2 speaker and, more importantly, generally require considerable speech data from each L2 speaker for training. To address these limitations, we propose Accentron, an approach that can generate accent-converted speech for arbitrary L2 speakers unseen during training. In the proposed approach, we first train a speaker-independent acoustic model on L1 corpora to extract bottleneck features that represent the linguistic content of utterances. Then, we develop a speaker encoder and an accent encoder to generate embedding vectors for the desired voice identity (L2 speaker’s) and accent (L1 accent), respectively. Lastly, we use a sequence-to-sequence model to transform bottleneck-features to Mel-spectrograms, conditioned on the L2 speaker embedding and the L1 accent embedding. We conducted experiments on the L2-ARCTIC corpus under two testing conditions: the standard FAC setting where test L2 speakers were seen during training, and a zero-shot FAC setting where test L2 speakers were unseen during training. Accentron achieves over 27% relative improvement in accentedness ratings compared to two state-of-the-art FAC systems in the standard FAC setting. More importantly, our results show that Accentron generalizes to the zero-shot FAC setting with no performance loss. Therefore, in practical use scenarios (e.g., computer-assisted pronunciation training software), Accentron can effectively avoid the need to adapt or retrain the model, which significantly reduces computations and the users’ waiting time.

Functionalization of optical waveguides with submicron coatings of zinc peroxide (ZnO2) and silica (SiO2) nanoparticles (NPs) is reported that enabled selective concentration of acetone vapors in the vicinity of the waveguide, boosting the sensitivity of a mid infrared (MIR) on-chip detector. Controlled thickness was achieved by introducing precise control of the substrate withdrawal speed to the layer-by-layer (LbL) deposition technique.

We previously developed a tablet-based speech therapy game called Apraxia World to address barriers to treatment and increase child motivation during therapy. In this study, we examined pronunciation improvements, child engagement over time, and caregiver evaluation performance while using our game. We recruited ten children to play Apraxia World at home during two four-week treatment blocks, separated by a two-week break; nine of ten have completed the protocol at time of writing. In the treatment blocks, children's utterances were evaluated either by caregivers or an automated pronunciation framework. Preliminary analysis suggests that children made significant therapy gains with Apraxia World, even though caregivers evaluated pronunciation leniently. We also collected a corpus of child speech for offline examination. We will conduct additional analysis once all participants complete the protocol.

Induced pluripotent stem cells (iPSCs) undergo unlimited self-renewal while maintaining their potential to differentiate into post-mitotic cells with an intact proteome. As such, iPSCs suppress the aggregation of polyQ-expanded huntingtin (HTT), the mutant protein underlying Huntington's disease (HD). Here we show that proteasome activity determines HTT levels, preventing polyQ-expanded aggregation in iPSCs from HD patients (HD-iPSCs). iPSCs exhibit high levels of UBR5, a ubiquitin ligase required for proteasomal degradation of both normal and mutant HTT. Conversely, loss of UBR5 increases HTT levels and triggers polyQ-expanded aggregation in HD-iPSCs. Moreover, UBR5 knockdown hastens polyQ-expanded aggregation and neurotoxicity in invertebrate models. Notably, UBR5 overexpression induces polyubiquitination and degradation of mutant HTT, reducing polyQ-expanded aggregates in HD-cell models. Besides HTT levels, intrinsic enhanced UBR5 expression determines global proteostasis of iPSCs preventing the aggregation of misfolded proteins ensued from normal metabolism. Thus, our findings indicate UBR5 as a modulator of super-vigilant proteostasis of iPSCs.

Working in an environment with constant interruptions is known to affect stress, but how do interruptions affect emotional expression? Emotional expression can have significant impact on interactions among coworkers. We analyzed the video of 26 participants who performed an essay task in a laboratory while receiving either continual email interruptions or receiving a single batch of email. Facial videos of the participants were run through a convolutional neural network to determine the emotional mix via decoding of facial expressions. Using a novel co-occurrence matrix analysis, we showed that with batched email, a neutral emotional state is dominant with sadness being a distant second, and with continual interruptions, this pattern is reversed, and sadness is mixed with fear. We discuss the implications of these results for how interruptions can impact employees' well-being and organizational climate.

Purpose: To assist in remote treatment, speech-language pathologists (SLPs) rely on mobile games, which though entertaining, lack feedback mechanisms. Games integrated with automatic speech recognition (ASR) offer a solution where speech productions control gameplay. We therefore performed a feasibility study to assess children's and SLPs' experiences towards speech-controlled games, game feature preferences and ASR accuracy. Method: Ten children with childhood apraxia of speech (CAS), six typically developing (TD) children and seven SLPs trialled five games and answered questionnaires. Researchers also compared the results of ASR to perceptual judgment. Result: Children and SLPs found speech-controlled games interesting and fun, despite ASR-human disagreements. They preferred games with rewards, challenge and multiple difficulty levels. Automatic speech recognition-human agreement was higher for SLPs than children, similar between TD and CAS and unaffected by CAS severity (77% TD, 75% CAS - incorrect; 51% TD, 47% CAS, 71% SLP - correct). Manual stop recording yielded higher agreement than automatic. Word length did not influence agreement. Conclusion: Children's and SLPs' positive responses towards speech-controlled games suggest that they can engage children in higher intensity practice. Our findings can guide future improvements to the ASR, recording methods and game features to improve the user experience and therapy adherence.

A systematic search and review of published studies was conducted on the use of automated speech analysis (ASA) tools for analysing and modifying speech of typically-developing children learning a foreign language and children with speech sound disorders to determine (i) types, attributes, and purposes of ASA tools being used; (ii) accuracy against human judgment; and (iii) performance as therapeutic tools.Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines were applied. Across nine databases, 32 articles published between January 2007 and December 2016 met inclusion criteria: (i) focussed on children's speech; (ii) tools used for speech analysis or modification; and (iii) reporting quantitative data on accuracy.Eighteen ASA tools were identified. These met the clinical threshold of 80% agreement with human judgment when used as predictors of intelligibility, impairment severity, or error category. Tool accuracy was typically 80% accuracy for words containing mispronunciations. ASA tools have been used effectively to improve to children's foreign language pronunciation.ASA tools show promise for automated analysis and modification of children's speech production within assessment and therapeutic applications. Further work is needed to train automated systems with larger samples of speech to increase accuracy for assessment and therapeutic feedback.

Optical character recognition (OCR) engines work poorly on texts published with premodern printing technologies. Engaging the key technological contributors from the IMPACT project, an earlier project attempting to solve the OCR problem for early modern and modern texts, the Early Modern OCR Project (eMOP) of Texas A8M received funding from the Andrew W. Mellon Foundation to improve OCR outputs for early modern texts from the Eighteenth Century Collections Online (ECCO) and Early English Books Online (EEBO) proprietary database products—or some 45 million pages. Added to print problems are the poor quality of the page images in these collections, which would be too time consuming and expensive to reimage. This article describes eMOP's attempts to OCR 307,000 documents digitized from microfilm to make our cultural heritage available for current and future researchers. We describe the reasoning behind our choices as we undertook the project based on other relevant studies; discoveries we made; the data and the system we developed for processing it; the software, algorithms, training procedures, and tools that we developed; and future directions that should be taken for further work in developing OCR engines for cultural heritage materials.

In 2014, 8.5% of people worldwide over the age of 18 had type 2 diabetes mellitus. It is estimated that, in recent years, the prevalence of this disease has increased mainly in low- and middle-income countries. Health behaviors are considered fundamental in the prevention, treatment, and control of this disease. The aim of this study was to determine if there is an association between health behaviors and satisfaction with family functioning in patients diagnosed with type 2 diabetes mellitus. This is an observational, transversal, analytical study. In the Family Medicine Unit No.33 in Chihuahua, Mexico, 297 patients with a diagnosis of type 2 diabetes mellitus were identified. Data was collected regarding sex, age and the Instrument for Measuring Lifestyles in Diabetics (IMEVID, for its acronym in Spanish) and the Family APGAR. We found that of the total number of patients, 91.6% presented some degree of unfavorable health behaviors, while 92.6% presented some type of family dysfunction. The categorical analysis positively correlated good family functionality with favorable health behaviors, Fisher's exact test p [Med-Science 2017; 6(4.000): 659-662]