教学课程资源库 更多>>

关键词： Hidden Markov models   Acoustics   Mathematical models   Feature extraction   Mel frequency cepstral coefficient   Computer science   Computational modeling   Pronunciation scoring   goodness of pronunciation   hidden Markov model-deep neural network   dynamic time warping   kaldi   speechocean762  

摘要： The current pronunciation scoring based on Goodness of Pronunciation (GOP) uses posterior probabilities of the Acoustic Models. Such algorithms suffer from generalization since they are utilized to determine a score metric for each phoneme rather than on the completeness or comparison with the ideal utterance of the words. This paper proposes a novel method to overcome such limitations by using combined scores of prosodic, fluency, completeness, and accuracy. This is achieved using context-aware GOP in conjugation with dynamic time warping (DTW) matching of the pitch contours of a weighted average of the context tokens found in the audio file that is rich in mispronounced phonemes. The proposed work gives flexibility in tuning the results according to different speech aspects based on a single hyperparameter. The results achieved are encouraging and have been validated on the speechocean762 dataset, where Automatic Speech Recognition (ASR) model has been trained on the Librispeech dataset. The resultant mean error of the proposed approach is 3.38% and the value of the correlation coefficient achieved is 0.652.

关键词： Computer Science   Engineering & Technology   MathML  

ISBN: (数字)9788770228961 ISBN: (纸本)9788770228619

摘要： This concise guide to mobile crowdsourcing and crowdsensing vulnerabilities and countermeasures walks readers through a series of examples, discussions, tables, initiative figures, and diagrams to present to them security and privacy foundations and applications. Discussed approaches help build intuition to apply these concepts to a broad range of system security domains toward dimensioning of next generations of mobiles crowdsensing applications. This book offers vigorous techniques as well as new insights for both beginners and seasoned professionals. It reflects on recent advances and research achievements. Technical topics discussed in the book include but are not limited to: • Risks affecting crowdsensing platforms • Spatio-temporal privacy of crowdsourced applications • Differential privacy for data mining crowdsourcing • Blockchain-based crowdsourcing • Secure wireless mobile crowdsensing. This book is accessible to readers in mobile computer/communication industries as well as academic staff and students in computer science, electrical engineering, telecommunication systems, business information systems, and crowdsourced mobile app developers.

关键词： Constraint satisfaction problems   Hrushovski-Fraïssé Classes   Omega-categoricity   Triangle-free graphs   Unary expansions  

摘要： In Chapter 2 we study the CSP of unary expansions of directed graphs. When we expand the simple structure (ℤ,succ) with one unary predicate U, its CSP (Constraint Satisfaction Problem) may vary in complexity. We find some sufficient conditions for its tractability, prove bounds on its complexity, and then generalize our results to more complicated structures. We also give a Karp-equivalent characterization of CSP(ℤ,succ, U)s. Next, fixing α ∈ (0, 1], we generalized the axiomatizations in to the class of hereditarily linearly sparse Kn-free graphs in Chapter 3, and made efforts towards connecting this with almost sure theories of Shelah-Spencer graphs, a type of random graphs, which may be of interest in theory of computer science. In Chapter 4 we study the constraint satisfaction problems of selected infinite relational structures. For α ∈ (0, 5/ 6 ], Kα := the class of hereditarily α-sparse graphs, Mα := the generic structure of Kα; Kα,Kn−free,Mα,Kn−free := the Kn-free subclass and its generic structure respectively, we prove multiple structural properties of graphs presenting the Hrushovski-Fraϊssé class ´ Kα,TF, strongly indicating that this class should have a finite homomorphic presentation when α is close to 5/6 from below, which would imply NP-completeness. More general results are explored along. We also study the complexity-theoretic implications of our findings on a relevant decision problem, namely the constraint satisfaction problem of the corresponding generic structure. Equations not Included.

关键词： Feature extraction   Training   Shape   Adaptation models   Task analysis   Computer science   Bridges   Domain adaptation   transfer learning   radial structure matching  

摘要： Domain adaptation methods reduce domain shift typically by learning domain-invariant features. Most existing methods are built on distribution matching, e.g., adversarial domain adaptation, which tends to corrupt feature discriminability. In this paper, we propose Discriminative Radial Domain Adaptation (DRDA) which bridges source and target domains via a shared radial structure. It's motivated by the observation that as the model is trained to be progressively discriminative, features of different categories expand outwards in different directions, forming a radial structure. We show that transferring such an inherently discriminative structure would enable to enhance feature transferability and discriminability simultaneously. Specifically, we represent each domain with a global anchor and each category a local anchor to form a radial structure and reduce domain shift via structure matching. It consists of two parts, namely isometric transformation to align the structure globally and local refinement to match each category. To enhance the discriminability of the structure, we further encourage samples to cluster close to the corresponding local anchors based on optimal-transport assignment. Extensively experimenting on multiple benchmarks, our method is shown to consistently outperforms state-of-the-art approaches on varied tasks, including the typical unsupervised domain adaptation, multi-source domain adaptation, domain-agnostic learning, and domain generalization.

摘要： Many introductory computer science (CS) courses are intended to address the increased demand for computer literacy and the development of cross-cutting concepts and practices of computational thinking (CT). Colleges and universities offer introductory CS courses every semester toward this end. The issue is centered on how to support CT learning in a large introductory CS course. To do it, an introductory CS course was designed to enrich CT by embedding short episodes of assessments to facilitate learning (AfL) and to support students’ self-regulated learning (SRL). This chapter begins by articulating the intricacies of the underlying processes of SRL, principles of AfL, and computational thinking practices in CS. Using the undergraduate introductory CS course referred to above, key practices of SRL and AfL will be illustrated to demonstrate how student learning of CS programming is structured with various online and offline supports. The chapter then concludes by appraising aspects of this CS course with practical suggestions for supporting student SRL and their learning of CS programming. © 2023 Wiley Periodicals LLC.

关键词： Carbon dioxide   Carbon footprint   Transportation   Behavioral sciences   Predictive analytics   Predictive models   Education   Information technology   information technology   predictive analytics   students  

摘要： Global warming poses a significant challenge to environmental sustainability due to the high greenhouse gas emissions originating from human activities. Among the various sectors contributing to these emissions, the education sector, particularly at the university level, plays a crucial role in generating carbon footprints. While several studies have examined carbon footprints within universities, few have focused on understanding the behavior and activities of individuals. Statistics show that students are the primary contributors to carbon emissions in higher education. Despite the availability of various methods for calculating carbon emissions, limited studies have utilized such data to predict future trends. Therefore, this study aimed to develop a predictive analytics model that leveraged students' learning activities as a significant factor to predict future trends in university carbon emissions. Institut Teknologi Bandung (ITB) was used as a case study, especially computer science-related majors. The carbon emission calculation utilized a formula that incorporated various emission sources, including electricity, transportation, and paper consumption. In the 2022/2023 academic year, ITB generated 612.8 tons of CO $_{2}\text{e}$ . The prediction modeling employed the SVR algorithm and utilized historical data, such as carbon emissions from the last 30 days, and external information, such as weather, event-related data, and university data. The model's performance was evaluated using metrics, yielding values of 129.41 for MAE, 158.03 for RMSE, 0.98 for R2, and 15.83 for MAPE. The results provided insights for universities to assess their carbon footprint and raised awareness among the academic community, supporting decision-making to optimize carbon emissions at the students' level.

关键词： Fisher information matrix  

摘要： With the technological advancement of the internet of things, the development of smart cities has become imperative for the economic, social, and environmental sustainability of the country. Digitalization of water resources management is one of the sustainable indicators for future residential space. Electronic water meters with intelligent functions of real-time data transmission and storage have been replacing traditional mechanical water meters. However, the failed electronic water meters returned from the actual venues have a similar water-entry problem shown in the area of the screw locks. According to the fault analysis, the problem attributes to the expansion-contraction effects caused by temperature and humidity, which not only affects the accuracy of signal transmission but also accelerates meters failure, making the product unable to meet the lifetime requirement and government regulations. Therefore, in this research, we conducted an accelerated life testing based on the Hallberg-Peck model for the electronic water-meters design improvement. We established step-by-step testing procedures and standards in action. The complete failure data were collected;they were evaluated, estimated, and inferred by R-Shiny Interactive app, which is a user-friendly application for practitioners to enjoy an accessible, interactive interface for dealing with the complicated reliability model and data inference. Finally, the results can be a valuable reference for practitioners to plan the next stage of design changes or the manufacturing process. © 2022, Chinese Society for Quality. All rights reserved.

关键词： Optoelectronic devices and components   Computer science  

摘要： Tensor analytics lays the mathematical basis for the prosperous promotion of multiway signal processing. To increase computing throughput, mainstream processors transform tensor convolutions into matrix multiplications to enhance the parallelism of computing. However, such order-reducing transformation produces data duplicates and consumes additional memory. Here, we propose an integrated photonic tensor flow processor (PTFP) without digitally duplicating the input data. It outputs the convolved tensor as the input tensor 'flows' through the processor. The hybrid manipulation of optical wavelengths, space dimensions, and time delay steps, enables the direct representation and processing of high-order tensors in the optical domain. In the proof-of-concept experiment, an integrated processor manipulating wavelengths and delay steps is implemented for demonstrating the key functionalities of PTFP. The multi-channel images and videos are processed at the modulation rate of 20 Gbaud. A convolutional neural network for video action recognition is demonstrated on the processor, which achieves an accuracy of 97.9%.

关键词： Curricula  

摘要： The words "Programming is the second literacy" were coined more than 40 years ago [13], but never came to life. The paper develops and details that old slogan by proposing that the mainstream mathematics education in schools should merge with education in computer science / programming. Of course, this means a deep structural reform of school mathematics education. We are not talking about adapting the 20th century mathematics to the 21st century—as it outlined in [10, 19], we mean the 21st century mathematics education for the 21st century mathematics. To the best of our knowledge, this paper is perhaps the first known attempt to start a proper feasibility study for this reform. The scope of the paper does not allow us to touch the delicate socio-political (and financial) sides of the reform, we are looking only at general curricular and didactic aspects and possible directions of the reform. In particular, we indicate approaches to development of a Domain Specifiic Language (DSL) as a basis for all programming aspects of a new *** Codes 97D99, 97C99 Copyright © 2022, The Authors. All rights reserved.

关键词： Knowledge graph   Scholarly domain   Scientific facts   Artificial intelligence  

摘要： Science communication has a number of bottlenecks that include the rising number of published research papers and its non-machine-accessible and document-based paradigm, which makes the ex-ploration, reading, and reuse of research outcomes rather inefficient. Recently, Knowledge Graphs (KG), i.e., semantic interlinked networks of entities, have been proposed as a new core technology to describe and curate scholarly information with the goal to make it machine readable and understandable. However, the main drawback of the use of such a technology is that researchers are asked to manually annotate their research papers and add their contributions within the KGs. To address this problem, in this paper we propose SCICERO, a novel KG generation approach that takes in input text from research articles and generates a KG of research entities. SCICERO uses Natural Language Processing techniques to parse the content of scientific papers to discover entities and relationships, exploits state-of-the-art Deep Learning Transformer models to make sense and validate extracted information, and uses Semantic Web best practices to formally represent the extracted entities and relationships, making the written content of research papers machine-actionable. SCICERO has been tested on a dataset of 6.7M papers about Computer Science generating a KG of about 10M entities. It has been evaluated on a manually generated gold standard of 3,600 triples that cover three Computer Science subdomains (Information Retrieval, Natural Language Processing, and Machine Learning) obtaining remarkable results.(c) 2022 Elsevier B.V. All rights reserved.