检索结果-内蒙古大学图书馆

Controllable Factuality in Document-Grounded Dialog Systems Using a Noisy Channel Model

学校读者我要写书评

暂无评论

arXiv 2022年

作者： Daheim, Nico Thulke, David Dugast, Christian Ney, Hermann Ubiquitous Knowledge Processing Lab Department of Computer Science Technical University of Darmstadt Germany Human Language Technology and Pattern Recognition RWTH Aachen University Germany AppTek GmbH Germany

In this work, we present a model for document-grounded response generation in dialog that is decomposed into two components according to Bayes' theorem. One component is a traditional ungrounded response generation model and the other component models the reconstruction of the grounding document based on the dialog context and generated response. We propose different approximate decoding schemes and evaluate our approach on multiple open-domain and task-oriented document-grounded dialog datasets. Our experiments show that the model is more factual in terms of automatic factuality metrics than the baseline model. Furthermore, we outline how introducing scaling factors between the components allows for controlling the tradeoff between factuality and fluency in the model output. Finally, we compare our approach to a recently proposed method to control factuality in grounded dialog, CTRL (Rashkin et al., 2021), and show that both approaches can be combined to achieve additional improvements. © 2022, CC BY.

关键词：

Unifying Input and Output Smoothing in Neural Machine Translation 28

学校读者我要写书评

暂无评论

Unifying Input and Output Smoothing in Neural Machine Transl...

28th International Conference on Computational Linguistics, COLING 2020

作者： Gao, Yingbo Liao, Baohao Ney, Hermann Human Language Technology and Pattern Recognition Group Computer Science Department RWTH Aachen University AachenD-52056 Germany

ISBN: (纸本)9781952148279

Soft contextualized data augmentation is a recent method that replaces one-hot representation of words with soft posterior distributions of an external language model, smoothing the input of neural machine translation systems. Label smoothing is another effective method that penalizes over-confident model outputs by discounting some probability mass from the true target word, smoothing the output of neural machine translation systems. Having the benefit of updating all word vectors in each optimization step and better regularizing the models, the two smoothing methods are shown to bring significant improvements in translation performance. In this work, we study how to best combine the methods and stack the improvements. Specifically, we vary the prior distributions to smooth with, the hyperparameters that control the smoothing strength, and the token selection procedures. We conduct extensive experiments on small datasets, evaluate the recipes on larger datasets, and examine the implications when back-translation is further used. Our results confirm cumulative improvements when input and output smoothing are used in combination, giving up to +1.9 BLEU scores on standard machine translation tasks and reveal reasons why these smoothing methods should be preferred. © 2020 COLING 2020 - 28th International Conference on Computational Linguistics, Proceedings of the Conference. All rights reserved.

关键词： Neural machine translation

Neural language Modeling for Named Entity Recognition 28

学校读者我要写书评

暂无评论

Neural Language Modeling for Named Entity Recognition

28th International Conference on Computational Linguistics, COLING 2020

作者： Lei, Zhihong Wang, Weiyue Dugast, Christian Ney, Hermann Apple Inc. Human Language Technology and Pattern Recognition Group Computer Science Department RWTH Aachen University Germany

ISBN: (纸本)9781952148279

Regardless of different word embedding and hidden layer structures of the neural architectures that are used in named entity recognition, a conditional random field layer is commonly used for the output. This work proposes to use a neural language model as an alternative to the conditional random field layer, which is more flexible for the size of the corpus. Experimental results show that the proposed system has a significant advantage in terms of training speed. © 2020 COLING 2020 - 28th International Conference on Computational Linguistics, Proceedings of the Conference. All rights reserved.

关键词： Modeling languages

Polarized Online Discourse on Abortion: Frames and Hostile Expressions among Liberals and Conservatives

学校读者我要写书评

暂无评论

arXiv 2023年

作者： Rao, Ashwin Chang, Rong-Ching Zhong, Qiankun Lerman, Kristina Wojcieszak, Magdalena University of Southern California Information Sciences Institute Marina del ReyCA United States Department of Computer Science University of California Davis United States Max Planck Institute for Human Development Germany Department of Communication University of California Davis United States Center for Excellence in Social Science University of Warsaw Poland

Abortion has been one of the most divisive issues in the United States. Yet, missing is comprehensive longitudinal evidence on how political divides on abortion are reflected in public discourse over time, on a national scale, and in response to key events before and after the overturn of Roe v Wade. We analyze a corpus of over 3.5M tweets related to abortion over the span of one year (January 2022 to January 2023) from over 1.1M users. We estimate users’ ideology and rely on state-of-the-art transformer-based classifiers to identify expressions of hostility and extract five prominent frames surrounding abortion. We use those data to examine (a) how prevalent were expressions of hostility (i.e., anger, toxic speech, insults, obscenities, and hate speech), (b) what frames liberals and conservatives used to articulate their positions on abortion, and (c) the prevalence of hostile expressions in liberals and conservative discussions of these frames. We show that liberals and conservatives largely mirrored each other’s use of hostile expressions: as liberals used more hostile rhetoric, so did conservatives, especially in response to key events. In addition, the two groups used distinct frames and discussed them in vastly distinct contexts, suggesting that liberals and conservatives have differing perspectives on abortion. Lastly, frames favored by one side provoked hostile reactions from the other: liberals use more hostile expressions when addressing religion, fetal personhood, and exceptions to abortion bans, whereas conservatives use more hostile language when addressing bodily autonomy and women’s health. This signals disrespect and derogation, which may further preclude understanding and exacerbate polarization. © 2023, CC BY.

关键词： Polarization

Analytics in Higher Education: Scoping the Landscape of Research in the Area 11

学校读者我要写书评

暂无评论

Analytics in Higher Education: Scoping the Landscape of Rese...

11th International Conference on System Modeling and Advancement in Research Trends, SMART 2022

作者： Kaul, Natashaa Deshpande, Amruta Mittal, Amit Jain, Vishal Symbiosis Center for Management and Human Resource Development Pune India India Symbiosis International University Pune India Indira School of Business Studies Pgdm Pune India Chitkara Business School Chitkara University Punjab India School of Engineering and Technology Sharda University Department of Computer Science and Engineering U.P. Greater Noida India

ISBN: (纸本)9781665487320

Understanding the research landscape of any domain is of primary importance at all stages of a concept's evolution. Research in the domain of analytics has increased exponentially in the past decade, analytics is a tool being applied to increase effectiveness and efficiency within all domains. Similarly, there are numerous research studies that have applied analytics to the domain of higher education. Assessing research activity is important for planning future research and identifying the way forward to make meaningful contributions. The objective of this paper is to examine the current research in this domain and to identify emerging areas of research to make meaningful contributions. Bibliometric analysis was applied using Scopus. Documents on analytics and higher education keywords in business, accounting and management and published in journals were considered for the analysis. Vos Viewer was used for bibliometric analysis in the area, to identify emerging research themes and to understand the scope of applying analytics in higher education. The search query found 1635 documents when using the keywords of 'analytics' and 'higher education'. This was further narrowed down to research in Business Management and Accounting and using Journal articles and conference articles, finally yielding 103 articles. After a primary read to remove articles misplaced in the group the final number of articles include for the bibliometric analysis came down to 98. There has been an increase in publications since 2018 with interest increasing post pandemic in the area. The research categories are Student outcomes and perspective, academician outcomes and perspective, design of learning systems, etc. Based on this, it is recommended that research should focus on theoretical aspects of applying analytics to HEIs, ethical issues in data collection, etc.. © 2022 IEEE.

关键词： Learning systems

Visualizations of Complex Sequences of Family-Infant Vocalizations Using Bag-of-Audio-Words Approach Based on Wav2vec 2.0 Features

学校读者我要写书评

暂无评论

arXiv 2022年

作者： Li, Jialu Hasegawa-Johnson, Mark McElwain, Nancy L. Department of Electrical and Computer Engineering University of Illinois United States Beckman Institute for Advanced Science and Technology University of Illinois United States Department of Human Development and Family Studies University of Illinois United States

In the U.S., approximately 15-17% of children 2-8 years of age are estimated to have at least one diagnosed mental, behavioral or developmental disorder. However, such disorders often go undiagnosed, and the ability to evaluate and treat disorders in the first years of life is limited. To analyze infant developmental changes, previous studies have shown advanced ML models excel at classifying infant and/or parent vocalizations collected using cell phone, video, or audio-only recording device like LENA. In this study, we pilot test the audio component of a new infant wearable multi-modal device that we have developed called LittleBeats (LB). LB audio pipeline is advanced in that it provides reliable labels for both speaker diarization and vocalization classification tasks, compared with other platforms that only record audio and/or provide speaker diarization labels. We leverage wav2vec 2.0 to obtain superior and more nuanced results with the LB family audio stream. We use a bag-of-audio-words method with wav2vec 2.0 features to create high-level visualizations to understand family-infant vocalization interactions. We demonstrate that our high-quality visualizations capture major types of family vocalization interactions, in categories indicative of mental, behavioral, and developmental health, for both labeled and unlabeled LB audio. Copyright © 2022, The Authors. All rights reserved.

关键词： Visualization

A Pilot Study of Sidewalk Equity in Seattle Using Crowdsourced Sidewalk Assessment Data

学校读者我要写书评

暂无评论

arXiv 2022年

作者： Li, Chu Orii, Lisa Saugstad, Michael Mooney, Stephen J. Eisenberg, Yochai Labbé, Delphine Hammel, Joy Froehlich, Jon E. Allen School of Computer Science University of Washington United States Epidemiology University of Washington United States Disability and Human Development Department University of Illinois Chicago United States

We examine the potential of using large-scale open crowdsourced sidewalk data from Project Sidewalk to study the distribution and condition of sidewalks in Seattle, WA. While potentially noisier than professionally gathered sidewalk datasets, crowdsourced data enables large, cross-regional studies that would be otherwise expensive and difficult to manage. As an initial case study, we examine spatial patterns of sidewalk quality in Seattle and their relationship to racial diversity, income level, built density, and transit modes. We close with a reflection on our approach, key limitations, and opportunities for future work. © 2022, CC BY.

关键词： Crowdsourcing

The Disconnect between Engineering Students' Desire to Discuss Racial Injustice in the Classroom and Faculty Anxieties

学校读者我要写书评

暂无评论

The Disconnect between Engineering Students' Desire to Discu...

2021 ASEE Virtual Annual Conference, ASEE 2021

作者： Hammond, Tracy Ray, Samantha Taele, Paul Thomas, Shawna Watson, Karan Stanley, Christine Polsley, Seth Institute for Engineering Education & Innovation Engineering Education Faculty Sketch Recognition Lab Department of Computer Science & Engineering Texas A&M University United States Engineering Education Faculty Department of Computer Science & Engineering Texas A&M University United States Engineering Education Faculty Department of Electrical and Computer Engineering Texas A&M University United States Engineering Education Faculty Department of Educational Administration & Human Development Texas A&M University United States

Protests against racial injustice have been increasing in the United States. Universities often rapidly respond to acts of injustice through public statements about their position to uphold the equality of all people. To gauge the desires and concerns around discussing events causing social unrest in engineering classrooms specifically, the engineering education faculty chair of a large university conducted discussions with both students and faculty regarding its place in their classrooms. This paper describes the emerging themes from survey responses using coding and grounded theory. Reactions from students and faculty were decidedly different. Most students stressed the importance of discussing such topics in class with their engineering faculty, while most faculty emphasized their concerns with doing so due to their lack of training to effectively handle such topics. This paper describes the evaluation of student and faculty responses and its implications for supporting diversity and inclusion in the engineering classroom. © American Society for Engineering Education, 2021

关键词： Students

Hybrid WDM and Optical-CDMA over Multi-Mode Fiber Transmission System based on Optical Vortex 5

学校读者我要写书评

暂无评论

Hybrid WDM and Optical-CDMA over Multi-Mode Fiber Transmissi...

5th International Conference on Electronic Design, ICED 2020

作者： Ghazi, Alaan Aljunid, S.A. Zulkarnain Syed Idrus, Syed Endut, R. Rashidi, C.B.M. Ali, N. Al-Dawoodi, Aras Fakhrudeen, Ahmed M. Fareed, Alaa Sharma, Teena School of Human Development and Techno-communication University Malaysia Perlis Malaysia School of Computer and Communication Engineering University Malaysia Perlis Malaysia Computer Science Department College of Computer Science and Information Technology Kirkuk University Iraq Networks Department College of Computer Science and Information Technology Kirkuk University Iraq Computer Technical Engineering Alkitab University Iraq School of Computing University Utara Malaysia Malaysia Centre of Excellence Universiti Malaysia Perlis Malaysia Universiti Malaysia Perlis Perlis Malaysia Department of Electronics and Communication Engineering Malviya National Institute of Technology Jaipur India

The optical vortex has recently attracted scholars to implement it in optical tweezers, microscopy, optical communications, quantum information processing, optical trapping, and laser machining. Optical vortex beam applied classically that can be transferred to the transverse amplitude of a heralded single-photon, and optical vortex possesses a helical wavefront and carries orbital angular momentum. In this study, Optical vortex is applied in optical-CDMA (optical code-division multiple-access) in conjunction with WDM (wavelength division multiplexing). This mechanism aims to increase the capacity and security in optical communication significantly. The implementation of Laguerre-Gaussian (LG) modes with optical vortex based on one dimension zero cross-correlation (ZCC) code shows that mode coupling reduces effectively. Consequently, a positive increase in channel performance and response. Accordingly, the LG modes based on the 1D-ZCC code are investigated and propagated over multi-mode fiber (MMF) based on an optical vortex, which also substantially reduces channel effects. Consequently, all these attributes combined will result in a hybrid WDM-Optical-CDMA with an optical vortex system over MMF. © 2021 Published under licence by IOP Publishing Ltd.

关键词： Optical tweezers