Browsing by Subject "Computer Science"

Now showing 1 - 6 of 6

Open Access
An energy stable C⁰ finite element scheme for a quasi-incompressible phase-field model of moving contact line with variable density
(Journal of Computational Physics, 2020-03-15) Shen, L; Huang, H; Lin, P; Song, Z; Xu, S
In this paper, we focus on modeling and simulation of two-phase flow problems with moving contact lines and variable density. A thermodynamically consistent phase-field model with general Navier boundary condition is developed based on the concept of quasi-incompressibility and the energy variational method. A mass conserving C0 finite element scheme is proposed to solve the PDE system. Energy stability is achieved at the fully discrete level. Various numerical results confirm that the proposed scheme for both P1 element and P2 element are energy stable.
Open Access
Automatic word count estimation from daylong child-centered recordings in various language environments using language-independent syllabification of speech
(Speech Communication, 2019-10-01) Räsänen, O; Seshadri, S; Karadayi, J; Riebling, E; Bunce, J; Cristia, A; Metze, F; Casillas, M; Rosemberg, C; Bergelson, E; Soderstrom, M
© 2019 The Authors Automatic word count estimation (WCE) from audio recordings can be used to quantify the amount of verbal communication in a recording environment. One key application of WCE is to measure language input heard by infants and toddlers in their natural environments, as captured by daylong recordings from microphones worn by the infants. Although WCE is nearly trivial for high-quality signals in high-resource languages, daylong recordings are substantially more challenging due to the unconstrained acoustic environments and the presence of near- and far-field speech. Moreover, many use cases of interest involve languages for which reliable ASR systems or even well-defined lexicons are not available. A good WCE system should also perform similarly for low- and high-resource languages in order to enable unbiased comparisons across different cultures and environments. Unfortunately, the current state-of-the-art solution, the LENA system, is based on proprietary software and has only been optimized for American English, limiting its applicability. In this paper, we build on existing work on WCE and present the steps we have taken towards a freely available system for WCE that can be adapted to different languages or dialects with a limited amount of orthographically transcribed speech data. Our system is based on language-independent syllabification of speech, followed by a language-dependent mapping from syllable counts (and a number of other acoustic features) to the corresponding word count estimates. We evaluate our system on samples from daylong infant recordings from six different corpora consisting of several languages and socioeconomic environments, all manually annotated with the same protocol to allow direct comparison. We compare a number of alternative techniques for the two key components in our system: speech activity detection and automatic syllabification of speech. As a result, we show that our system can reach relatively consistent WCE accuracy across multiple corpora and languages (with some limitations). In addition, the system outperforms LENA on three of the four corpora consisting of different varieties of English. We also demonstrate how an automatic neural network-based syllabifier, when trained on multiple languages, generalizes well to novel languages beyond the training data, outperforming two previously proposed unsupervised syllabifiers as a feature extractor for WCE.
Open Access
Behavior of different numerical schemes for random genetic drift
(BIT Numerical Mathematics, 2019-09-01) Xu, S; Chen, M; Liu, C; Zhang, R; Yue, X
In the problem of random genetic drift, the probability density of one gene is governed by a degenerated convection-dominated diffusion equation. Dirac singularities will always be developed at boundary points as time evolves, which is known as the fixation phenomenon in genetic evolution. Three finite volume methods: FVM1-3, one central difference method: FDM1 and three finite element methods: FEM1-3 are considered. These methods lead to different equilibrium states after a long time. It is shown that only schemes FVM3 and FEM3, which are the same, preserve probability, expectation and positiveness and predict the correct probability of fixation. FVM1-2 wrongly predict the probability of fixation due to their intrinsic viscosity, even though they are unconditionally stable. Contrarily, FDM1 and FEM1-2 introduce different anti-diffusion terms, which make them unstable and fail to preserve positiveness.
Open Access
Computer aided restoration of handwritten character strokes
(CAD Computer Aided Design, 2017-08-01) Sober, B; Levin, D
© 2017 Elsevier Ltd This work suggests a new variational approach to the task of computer aided segmentation and restoration of incomplete characters, residing in a highly noisy document image. We model character strokes as the movement of a pen with a varying radius. Following this model, in order to fit the digital image, a cubic spline representation is being utilized to perform gradient descent steps, while maintaining interpolation at some initial (manually sampled) points. The proposed algorithm was used in the process of restoring approximately 1000 ancient Hebrew characters (dating to ca. 8th–7th century BCE), some of which are presented herein and show that the algorithm yields plausible results when applied on deteriorated documents.
Open Access
Reduced-order deep learning for flow dynamics. The interplay between deep learning and model reduction
(Journal of Computational Physics, 2020-01) Wang, Min; Cheung, Siu Wun; Leung, Wing Tat; Chung, Eric T; Efendiev, Yalchin; Wheeler, Mary
Open Access
The Press and Peace
(2024-05-10) Bussey, Jakobe
This study utilizes state-of-the-art BERT (Bidirectional Encoder Representations from Transformers) models to perform sentiment analysis on Wall Street Journal and New York Times articles about the Iraq War published between 2002 and 2012 and further categorize them using advanced unsupervised machine learning techniques. By utilizing statistical analysis and quartic regression models, this paper concludes that the two newspapers report on the Iraq War differently, with both exhibiting a predominantly negative-neutral tone overall. Additionally, the analysis reveals significant fluctuations in negativity from both outlets over time as the war progresses. Furthermore, this study examines the objectivity of reporting between editorial and non-editorial articles, finding that non-editorials tend to report more objectively, and the neutrality of editorials remains relatively constant while the objectivity of non-editorials fluctuates in response to war events. Finally, the paper investigates variations in sentiment across different topics, uncovering substantial variations in positive, neutral, and negative sentiments across topics and their evolution over time.