Pub Date : 2022-08-19DOI: 10.48550/arXiv.2208.09299
Rebecca M. C. Taylor, J. D. Preez
Variational Bayes (VB) applied to latent Dirichlet allocation (LDA) has become the most popular algorithm for aspect modeling. While sufficiently successful in text topic extraction from large corpora, VB is less successful in identifying aspects in the presence of limited data. We present a novel variational message passing algorithm as applied to Latent Dirichlet Allocation (LDA) and compare it with the gold standard VB and collapsed Gibbs sampling. In situations where marginalisation leads to non-conjugate messages, we use ideas from sampling to derive approximate update equations. In cases where conjugacy holds, Loopy Belief update (LBU) (also known as Lauritzen-Spiegelhalter) is used. Our algorithm, ALBU (approximate LBU), has strong similarities with Variational Message Passing (VMP) (which is the message passing variant of VB). To compare the performance of the algorithms in the presence of limited data, we use data sets consisting of tweets and news groups. Using coherence measures we show that ALBU learns latent distributions more accurately than does VB, especially for smaller data sets.
{"title":"SimLDA: A tool for topic model evaluation","authors":"Rebecca M. C. Taylor, J. D. Preez","doi":"10.48550/arXiv.2208.09299","DOIUrl":"https://doi.org/10.48550/arXiv.2208.09299","url":null,"abstract":"Variational Bayes (VB) applied to latent Dirichlet allocation (LDA) has become the most popular algorithm for aspect modeling. While sufficiently successful in text topic extraction from large corpora, VB is less successful in identifying aspects in the presence of limited data. We present a novel variational message passing algorithm as applied to Latent Dirichlet Allocation (LDA) and compare it with the gold standard VB and collapsed Gibbs sampling. In situations where marginalisation leads to non-conjugate messages, we use ideas from sampling to derive approximate update equations. In cases where conjugacy holds, Loopy Belief update (LBU) (also known as Lauritzen-Spiegelhalter) is used. Our algorithm, ALBU (approximate LBU), has strong similarities with Variational Message Passing (VMP) (which is the message passing variant of VB). To compare the performance of the algorithms in the presence of limited data, we use data sets consisting of tweets and news groups. Using coherence measures we show that ALBU learns latent distributions more accurately than does VB, especially for smaller data sets.","PeriodicalId":93362,"journal":{"name":"Proceedings of the Future Technologies Conference (FTC) 2020. Future Technologies Conference (2020 : Online)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2022-08-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"77342831","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-06-21DOI: 10.48550/arXiv.2206.10389
T. Yamakami
The concept of space-bounded computability has become significantly important in handling vast data sets on memory-limited computing devices. To replenish the existing short list of NL-complete problems whose instance sizes are dictated by log-space size parameters, we propose new additions obtained directly from natural parameterizations of three typical NP-complete problems -- the vertex cover problem, the exact cover by 3-sets problem, and the 3-dimensional matching problem. With appropriate restrictions imposed on their instances, the proposed decision problems parameterized by appropriate size parameters are proven to be equivalent in computational complexity to either the parameterized $3$-bounded 2CNF Boolean formula satisfiability problem or the parameterized degree-$3$ directed $s$-$t$ connectivity problem by ``short'' logarithmic-space reductions. Under the assumption of the linear space hypothesis, furthermore, none of the proposed problems can be solved in polynomial time if the memory usage is limited to sub-linear space.
{"title":"Parameterized-NL Completeness of Combinatorial Problems by Short Logarithmic-Space Reductions and Immediate Consequences of the Linear Space Hypothesis","authors":"T. Yamakami","doi":"10.48550/arXiv.2206.10389","DOIUrl":"https://doi.org/10.48550/arXiv.2206.10389","url":null,"abstract":"The concept of space-bounded computability has become significantly important in handling vast data sets on memory-limited computing devices. To replenish the existing short list of NL-complete problems whose instance sizes are dictated by log-space size parameters, we propose new additions obtained directly from natural parameterizations of three typical NP-complete problems -- the vertex cover problem, the exact cover by 3-sets problem, and the 3-dimensional matching problem. With appropriate restrictions imposed on their instances, the proposed decision problems parameterized by appropriate size parameters are proven to be equivalent in computational complexity to either the parameterized $3$-bounded 2CNF Boolean formula satisfiability problem or the parameterized degree-$3$ directed $s$-$t$ connectivity problem by ``short'' logarithmic-space reductions. Under the assumption of the linear space hypothesis, furthermore, none of the proposed problems can be solved in polynomial time if the memory usage is limited to sub-linear space.","PeriodicalId":93362,"journal":{"name":"Proceedings of the Future Technologies Conference (FTC) 2020. Future Technologies Conference (2020 : Online)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2022-06-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"81275604","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-05-06DOI: 10.48550/arXiv.2205.03330
Sreejith Gopinath, Aspen Olmsted
The rapid advancement of technology has resulted in advanced camera capabilities coming to smaller form factors with improved energy efficiency. These improvements have led to more efficient and capable cameras on mobile devices like mobile phones, tablets, and even eyeglasses. Using these unobtrusive cameras, users can capture photographs and videos of almost any location where they have physical access. Unfortunately, the proliferation of highly compact cameras has threatened the privacy rights of individuals and even entire nations and governments. For example, governments may not want photographs or videos of sensitive installations or locations like airside operations of military bases or the inner areas of nuclear power plants to be captured for unapproved uses. In addition, solutions that obfuscate images in post-processing are subject to threats that could siphon unprocessed data. Our work proposes a Global Positioning System-based approach to restrict the ability of smart cameras to capture and store images of sensitive areas.
{"title":"Safeguarding National Security Interests Utilizing Location-Aware Camera Devices","authors":"Sreejith Gopinath, Aspen Olmsted","doi":"10.48550/arXiv.2205.03330","DOIUrl":"https://doi.org/10.48550/arXiv.2205.03330","url":null,"abstract":"The rapid advancement of technology has resulted in advanced camera capabilities coming to smaller form factors with improved energy efficiency. These improvements have led to more efficient and capable cameras on mobile devices like mobile phones, tablets, and even eyeglasses. Using these unobtrusive cameras, users can capture photographs and videos of almost any location where they have physical access. Unfortunately, the proliferation of highly compact cameras has threatened the privacy rights of individuals and even entire nations and governments. For example, governments may not want photographs or videos of sensitive installations or locations like airside operations of military bases or the inner areas of nuclear power plants to be captured for unapproved uses. In addition, solutions that obfuscate images in post-processing are subject to threats that could siphon unprocessed data. Our work proposes a Global Positioning System-based approach to restrict the ability of smart cameras to capture and store images of sensitive areas.","PeriodicalId":93362,"journal":{"name":"Proceedings of the Future Technologies Conference (FTC) 2020. Future Technologies Conference (2020 : Online)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2022-05-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"83485554","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-04-19DOI: 10.48550/arXiv.2204.08791
Hannes Fassold
A good optical flow estimation is crucial in many video analysis and restoration algorithms employed in application fields like media industry, industrial inspection and automotive. In this work, we investigate how well optical flow algorithms perform qualitatively when integrated into a state of the art video denoising algorithm. Both classic optical flow algorithms (e.g. TV-L1) as well as recent deep learning based algorithm (like RAFT or BMBC) will be taken into account. For the qualitative investigation, we will employ realistic content with challenging characteristic (noisy content, large motion etc.) instead of the standard images used in most publications.
{"title":"A qualitative investigation of optical flow algorithms for video denoising","authors":"Hannes Fassold","doi":"10.48550/arXiv.2204.08791","DOIUrl":"https://doi.org/10.48550/arXiv.2204.08791","url":null,"abstract":"A good optical flow estimation is crucial in many video analysis and restoration algorithms employed in application fields like media industry, industrial inspection and automotive. In this work, we investigate how well optical flow algorithms perform qualitatively when integrated into a state of the art video denoising algorithm. Both classic optical flow algorithms (e.g. TV-L1) as well as recent deep learning based algorithm (like RAFT or BMBC) will be taken into account. For the qualitative investigation, we will employ realistic content with challenging characteristic (noisy content, large motion etc.) instead of the standard images used in most publications.","PeriodicalId":93362,"journal":{"name":"Proceedings of the Future Technologies Conference (FTC) 2020. Future Technologies Conference (2020 : Online)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2022-04-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"80867536","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-03-24DOI: 10.1007/978-3-031-18461-1_13
Alex Zhang, N. Li
{"title":"A Two-Stage Federated Transfer Learning Framework in Medical Images Classification on Limited Data: A COVID-19 Case Study","authors":"Alex Zhang, N. Li","doi":"10.1007/978-3-031-18461-1_13","DOIUrl":"https://doi.org/10.1007/978-3-031-18461-1_13","url":null,"abstract":"","PeriodicalId":93362,"journal":{"name":"Proceedings of the Future Technologies Conference (FTC) 2020. Future Technologies Conference (2020 : Online)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2022-03-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"78141259","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-01-01DOI: 10.1007/978-3-031-18458-1_8
P. Boulanger
{"title":"Is Free-Viewpoint Video the Future of Online TV?","authors":"P. Boulanger","doi":"10.1007/978-3-031-18458-1_8","DOIUrl":"https://doi.org/10.1007/978-3-031-18458-1_8","url":null,"abstract":"","PeriodicalId":93362,"journal":{"name":"Proceedings of the Future Technologies Conference (FTC) 2020. Future Technologies Conference (2020 : Online)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2022-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"75856949","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-01-01DOI: 10.1007/978-3-031-18344-7_54
E. M. Gertis, Y. Liang
{"title":"Design and Implementation of an Automatic Word Match Generator","authors":"E. M. Gertis, Y. Liang","doi":"10.1007/978-3-031-18344-7_54","DOIUrl":"https://doi.org/10.1007/978-3-031-18344-7_54","url":null,"abstract":"","PeriodicalId":93362,"journal":{"name":"Proceedings of the Future Technologies Conference (FTC) 2020. Future Technologies Conference (2020 : Online)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2022-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"74647584","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-01-01DOI: 10.1007/978-3-031-18458-1_33
Abdelfateh Rahmouni, S. Oukid, Yasmina Ghebghoub
{"title":"Upgrade SecSLA Using Security as a Service Based on Knowledge Graph","authors":"Abdelfateh Rahmouni, S. Oukid, Yasmina Ghebghoub","doi":"10.1007/978-3-031-18458-1_33","DOIUrl":"https://doi.org/10.1007/978-3-031-18458-1_33","url":null,"abstract":"","PeriodicalId":93362,"journal":{"name":"Proceedings of the Future Technologies Conference (FTC) 2020. Future Technologies Conference (2020 : Online)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2022-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"75421154","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：