Pub Date : 2022-11-29DOI: 10.1109/SNAMS58071.2022.10062827
W. Farhan, Jamil Razmak
Social networks used for health-related services can either support or hinder the adoption of Canadian ehealth strategies. Assessing the Canadian usage of ehealth services, including social networks, was the main objective of the present study, which used secondary data from the Canadian Medical Association and the Canadian Health Infoway. Responses from 2,071 physicians and 1,017 patients were analysed using SPSS. The results indicated that 99.8% of Canadian physicians did not use social networks, while 14% of Canadian patients did report using them. Using Chi-square analysis, we found significant differences in social network use among patients according to age, gender, and employment status. Effective and careful engagement in social networks encouraged by Canadian health policy makers for both physicians and patients, the spread of correct and trusted health-related information, and the provision of trusted communication channels on social network sites benefit both public and private healthcare institutions, and can support and boost the achievements of the desired ehealth goals.
{"title":"Social networking as an eHealth tool: Its usage by Canadian physicians and patients","authors":"W. Farhan, Jamil Razmak","doi":"10.1109/SNAMS58071.2022.10062827","DOIUrl":"https://doi.org/10.1109/SNAMS58071.2022.10062827","url":null,"abstract":"Social networks used for health-related services can either support or hinder the adoption of Canadian ehealth strategies. Assessing the Canadian usage of ehealth services, including social networks, was the main objective of the present study, which used secondary data from the Canadian Medical Association and the Canadian Health Infoway. Responses from 2,071 physicians and 1,017 patients were analysed using SPSS. The results indicated that 99.8% of Canadian physicians did not use social networks, while 14% of Canadian patients did report using them. Using Chi-square analysis, we found significant differences in social network use among patients according to age, gender, and employment status. Effective and careful engagement in social networks encouraged by Canadian health policy makers for both physicians and patients, the spread of correct and trusted health-related information, and the provision of trusted communication channels on social network sites benefit both public and private healthcare institutions, and can support and boost the achievements of the desired ehealth goals.","PeriodicalId":371668,"journal":{"name":"2022 Ninth International Conference on Social Networks Analysis, Management and Security (SNAMS)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2022-11-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132166179","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 : 2022-11-29DOI: 10.1109/SNAMS58071.2022.10062585
Khaled Balhaf, Omar A. Darwish, Emad Rawashdeh, Mohammad Abu Awad, Dirar A. Darweesh, Yahya M. Tashtoush, Saif Rawashdeh
Recently, media and business companies are utilizing social media to reach a large set of users to maximize the amount of gained profit. Actually, these companies are looking for the best ways to satisfy their user's requirements. It is very difficult to understand these requirements because of the large set of users on social media like Twitter. For this reason, the goal of our research project is to build a classifier that can detect Arabian trends among Gulf area Twitter users. The new built classifier can assist these companies to deliver the convenient products and media contents like photos and videos according to users' trends. By using our own designed Java-based tool, we have collected a significant dataset of tweets. Also, two experiments of tweet classification have been implemented to compare the effects of balanced and imbalanced training data and to measure the effect of data size on the accuracy of classifiers. In both experiments, Support Vector Machine (SVM), K-Nearest Neighbors (KNN) and Naïve Bayes algorithms are used as classifiers. The first experiment uses small, imbalanced data sets and four classes of data, which are Sport, Politics, Islam and Culture. The Light and Root Stemmers were used with each classifier. The best outcome achieved in our research project by utilizing a Naïve Bayes algorithm with the Light Stemmer technique. It achieved an accuracy reaching 76.27%. In the second experiment, we used a balanced large data set with the same classifiers. In addition, we have added one more class to the new data set which is Economics. The experimental results showed that the best accuracy (81.17%) is obtained by using SVM with the Light Stemmer method. The Light Stemmer achieved the best outcomes for all classifiers since almost all of the tweets were written in dialects.
{"title":"Classifying Arabian Gulf Tweets to Detect People's Trends: A case study","authors":"Khaled Balhaf, Omar A. Darwish, Emad Rawashdeh, Mohammad Abu Awad, Dirar A. Darweesh, Yahya M. Tashtoush, Saif Rawashdeh","doi":"10.1109/SNAMS58071.2022.10062585","DOIUrl":"https://doi.org/10.1109/SNAMS58071.2022.10062585","url":null,"abstract":"Recently, media and business companies are utilizing social media to reach a large set of users to maximize the amount of gained profit. Actually, these companies are looking for the best ways to satisfy their user's requirements. It is very difficult to understand these requirements because of the large set of users on social media like Twitter. For this reason, the goal of our research project is to build a classifier that can detect Arabian trends among Gulf area Twitter users. The new built classifier can assist these companies to deliver the convenient products and media contents like photos and videos according to users' trends. By using our own designed Java-based tool, we have collected a significant dataset of tweets. Also, two experiments of tweet classification have been implemented to compare the effects of balanced and imbalanced training data and to measure the effect of data size on the accuracy of classifiers. In both experiments, Support Vector Machine (SVM), K-Nearest Neighbors (KNN) and Naïve Bayes algorithms are used as classifiers. The first experiment uses small, imbalanced data sets and four classes of data, which are Sport, Politics, Islam and Culture. The Light and Root Stemmers were used with each classifier. The best outcome achieved in our research project by utilizing a Naïve Bayes algorithm with the Light Stemmer technique. It achieved an accuracy reaching 76.27%. In the second experiment, we used a balanced large data set with the same classifiers. In addition, we have added one more class to the new data set which is Economics. The experimental results showed that the best accuracy (81.17%) is obtained by using SVM with the Light Stemmer method. The Light Stemmer achieved the best outcomes for all classifiers since almost all of the tweets were written in dialects.","PeriodicalId":371668,"journal":{"name":"2022 Ninth International Conference on Social Networks Analysis, Management and Security (SNAMS)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2022-11-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114201832","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 : 2022-11-29DOI: 10.1109/SNAMS58071.2022.10062744
Yuhang Zhang, Q. Zhang, Man Jiang, Jiangtao Su
Deep learning has been widely applied for jobs involving face inpainting, however, there are usually some problems, such as incoherent inpainting edges, lack of diversity of generated images and other problems. In order to get more feature information and improve the inpainting effect, we therefore propose a Generative Adversarial Network of Skip Connection (SCGAN), which connects the encoder layers and the decoder layers by skip connection in the generator. The coherence and consistency of the image inpainting edges are improved, and the finer features of the image inpainting are refined, simultaneously using the discriminator's local and global double discriminators model. We also employ WGAN-GP loss to enhance model stability during training, prevent model collapse, and increase the variety of inpainting face images. Finally, experiments on the CelebA dataset and the LFW dataset are performed, and the model's performance is assessed using the PSNR and SSIM indices. Our model's face image inpainting is more realistic and coherent than that of other models, and the model training is more reliable.
{"title":"SCGAN: Generative Adversarial Networks of Skip Connection for Face Image Inpainting","authors":"Yuhang Zhang, Q. Zhang, Man Jiang, Jiangtao Su","doi":"10.1109/SNAMS58071.2022.10062744","DOIUrl":"https://doi.org/10.1109/SNAMS58071.2022.10062744","url":null,"abstract":"Deep learning has been widely applied for jobs involving face inpainting, however, there are usually some problems, such as incoherent inpainting edges, lack of diversity of generated images and other problems. In order to get more feature information and improve the inpainting effect, we therefore propose a Generative Adversarial Network of Skip Connection (SCGAN), which connects the encoder layers and the decoder layers by skip connection in the generator. The coherence and consistency of the image inpainting edges are improved, and the finer features of the image inpainting are refined, simultaneously using the discriminator's local and global double discriminators model. We also employ WGAN-GP loss to enhance model stability during training, prevent model collapse, and increase the variety of inpainting face images. Finally, experiments on the CelebA dataset and the LFW dataset are performed, and the model's performance is assessed using the PSNR and SSIM indices. Our model's face image inpainting is more realistic and coherent than that of other models, and the model training is more reliable.","PeriodicalId":371668,"journal":{"name":"2022 Ninth International Conference on Social Networks Analysis, Management and Security (SNAMS)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2022-11-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128137373","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 : 2022-11-29DOI: 10.1109/SNAMS58071.2022.10062594
Johannes Schoder, H. M. Bücker, André Bötticher, Ronja M. Karmann
To compare different types of music recommender systems, datasets are necessary that offer a combination of diverse features. We propose MusicDress, a novel dataset covering four different elements of music: timbre, rhythm, melody, and harmony. The dataset extends to lyrics and user data by linking to publicly available data sources. It comprises features of 2,136 individual songs and enables the comparison of hybrid recommender systems that combine content-based, context-based, and collaborative filtering approaches.
{"title":"MusicDress: A Heterogeneous Dataset for Comparing Music Recommender Systems","authors":"Johannes Schoder, H. M. Bücker, André Bötticher, Ronja M. Karmann","doi":"10.1109/SNAMS58071.2022.10062594","DOIUrl":"https://doi.org/10.1109/SNAMS58071.2022.10062594","url":null,"abstract":"To compare different types of music recommender systems, datasets are necessary that offer a combination of diverse features. We propose MusicDress, a novel dataset covering four different elements of music: timbre, rhythm, melody, and harmony. The dataset extends to lyrics and user data by linking to publicly available data sources. It comprises features of 2,136 individual songs and enables the comparison of hybrid recommender systems that combine content-based, context-based, and collaborative filtering approaches.","PeriodicalId":371668,"journal":{"name":"2022 Ninth International Conference on Social Networks Analysis, Management and Security (SNAMS)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2022-11-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128407231","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 : 2022-11-29DOI: 10.1109/SNAMS58071.2022.10062544
R. Ivanova, Stefan Sobernig, Mark Strembeck
We report on an experiment that used ten different machines running on a standardized cloud platform in five different geographical locations around the globe (Frankfurt/Germany, Mumbai/India, Sydney/Australia, Seoul/South Korea, Virginia/USA) to collect datasets using Twitter's public free-of-charge API. Each of the ten machines extracted the tweets at the exact same time and using the exact same Twitter API parameters. We found that the characteristics of the datasets collected in different locations vary considerably, potentially affecting any analysis performed on such location-biased data. For example, the number of exactly identical tweets (i.e. all 90 metadata attributes of the tweets are the same for all ten machines) lays only between 0.15% and 20%. Based on these findings, we derive recommendations on how to mitigate the location-bias in practice.
{"title":"Does geographical location have an impact on data samples extracted from Twitter?","authors":"R. Ivanova, Stefan Sobernig, Mark Strembeck","doi":"10.1109/SNAMS58071.2022.10062544","DOIUrl":"https://doi.org/10.1109/SNAMS58071.2022.10062544","url":null,"abstract":"We report on an experiment that used ten different machines running on a standardized cloud platform in five different geographical locations around the globe (Frankfurt/Germany, Mumbai/India, Sydney/Australia, Seoul/South Korea, Virginia/USA) to collect datasets using Twitter's public free-of-charge API. Each of the ten machines extracted the tweets at the exact same time and using the exact same Twitter API parameters. We found that the characteristics of the datasets collected in different locations vary considerably, potentially affecting any analysis performed on such location-biased data. For example, the number of exactly identical tweets (i.e. all 90 metadata attributes of the tweets are the same for all ten machines) lays only between 0.15% and 20%. Based on these findings, we derive recommendations on how to mitigate the location-bias in practice.","PeriodicalId":371668,"journal":{"name":"2022 Ninth International Conference on Social Networks Analysis, Management and Security (SNAMS)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2022-11-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"134213122","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 : 2022-11-29DOI: 10.1109/SNAMS58071.2022.10062524
Abdelrahman Shalaby, Gehad Mohamed Abdelaziz, M. Kandeel
This article discusses using Artificial Intelligence in the context of law, specifically online arbitration. The article starts with some theoretical basis discussing the concept of AI, its development, and its advantages and disadvantages. The practical part of this article focuses on the main concerns when using AI in online arbitration. The first concern is using AI to assist with the arbitrator's selection. The second concern is using AI to issue an arbitral award. The article concludes that AI can positively impact online arbitration while taking into consideration that human intervention is necessary in some cases.
{"title":"Using Artificial Intelligence to Resolve Disputes through Online Arbitration","authors":"Abdelrahman Shalaby, Gehad Mohamed Abdelaziz, M. Kandeel","doi":"10.1109/SNAMS58071.2022.10062524","DOIUrl":"https://doi.org/10.1109/SNAMS58071.2022.10062524","url":null,"abstract":"This article discusses using Artificial Intelligence in the context of law, specifically online arbitration. The article starts with some theoretical basis discussing the concept of AI, its development, and its advantages and disadvantages. The practical part of this article focuses on the main concerns when using AI in online arbitration. The first concern is using AI to assist with the arbitrator's selection. The second concern is using AI to issue an arbitral award. The article concludes that AI can positively impact online arbitration while taking into consideration that human intervention is necessary in some cases.","PeriodicalId":371668,"journal":{"name":"2022 Ninth International Conference on Social Networks Analysis, Management and Security (SNAMS)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2022-11-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125556978","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 : 2022-11-29DOI: 10.1109/icdec.2016.7563134
{"title":"Committees Members","authors":"","doi":"10.1109/icdec.2016.7563134","DOIUrl":"https://doi.org/10.1109/icdec.2016.7563134","url":null,"abstract":"","PeriodicalId":371668,"journal":{"name":"2022 Ninth International Conference on Social Networks Analysis, Management and Security (SNAMS)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2022-11-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124043869","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 : 2022-11-29DOI: 10.1109/SNAMS58071.2022.10062551
F. M. F. Saboune
Advertisers and media owners are becoming more dependent on advertising, when the traditional way is losing its effect, they are eagerly looking for alternative ways to reach consumers, and keeping their brand trust, promise, and image at the top, and to generate more profit. Although there are many ways companies and brands alike can seek to make money off of social media, the traditional platforms are becoming obsolete. Gone are the days of having the logo, a photo, and an attractive slogan sufficient for attracting new customers. It can be said that advertising relies heavily on good aesthetics and application of the brand image, however applying that correctly to the target audience effectively can enhance and produce better outcomes, it's about how a business achieves success, scale worldwide, persuade consumers to buy into it, and reach its goals and profits. To create an effective and impactful advertisement in our modern world, companies must turn to technology; For technology is the enabler of a wider selection of tools that help businesses make better data-driven and informative decisions. This can be done by collecting data through third parties, internal or external surveys, and polls to enhance and improve their product delivery and message. With the use of data, algorithms can be then created to understand the consumer's behavior, therefore, enabling a business to better target its user-base based on their interests, search patterns, and online behavior. Moreover, the collected data from the mentioned tools can be then transformed into resourceful decisions to help the business better understand its user-base, therefore choosing from a plethora of options such as using influencers to advertise their product on live-video streams, or the traditional bloggers, and the many content creators available. Web3 is on the horizon, with the accelerative pace the metaverse took over the internet, brands turned to the newly found virtual space to promote themselves. This is a stepping stone into what is to come next as it taps into our day-to-day lives, economy, and social interactions [16]. The main component of web3 is the ability to build on it and interact with the web like never before and securely allowing users to maintain their privacy. Users are able to buy, maintain and sell virtual assets on the blockchain which unleashes a whole new era of using digital goods paved by the improvement of Virtual and Augmented realities known as VR and AR [1]. This has been proven by the famously known “Metaverse” world where companies such as Nintendo, Sony, and other gaming brands as well as celebrities such as Snoop Dogg rushed to create their own virtual theme parks and sold “plot of lands” in the said universe on the blockchain known as NFTs (Non-Fungible Tokens). This has created a ripple effect that continued in other virtually made assets, mainly video games, and lastly “The Metaverse” that “Meta” formerly known as “Facebook” is building upon.
{"title":"Virtual Reality in Social media marketing will be the new model of advertising and monetization","authors":"F. M. F. Saboune","doi":"10.1109/SNAMS58071.2022.10062551","DOIUrl":"https://doi.org/10.1109/SNAMS58071.2022.10062551","url":null,"abstract":"Advertisers and media owners are becoming more dependent on advertising, when the traditional way is losing its effect, they are eagerly looking for alternative ways to reach consumers, and keeping their brand trust, promise, and image at the top, and to generate more profit. Although there are many ways companies and brands alike can seek to make money off of social media, the traditional platforms are becoming obsolete. Gone are the days of having the logo, a photo, and an attractive slogan sufficient for attracting new customers. It can be said that advertising relies heavily on good aesthetics and application of the brand image, however applying that correctly to the target audience effectively can enhance and produce better outcomes, it's about how a business achieves success, scale worldwide, persuade consumers to buy into it, and reach its goals and profits. To create an effective and impactful advertisement in our modern world, companies must turn to technology; For technology is the enabler of a wider selection of tools that help businesses make better data-driven and informative decisions. This can be done by collecting data through third parties, internal or external surveys, and polls to enhance and improve their product delivery and message. With the use of data, algorithms can be then created to understand the consumer's behavior, therefore, enabling a business to better target its user-base based on their interests, search patterns, and online behavior. Moreover, the collected data from the mentioned tools can be then transformed into resourceful decisions to help the business better understand its user-base, therefore choosing from a plethora of options such as using influencers to advertise their product on live-video streams, or the traditional bloggers, and the many content creators available. Web3 is on the horizon, with the accelerative pace the metaverse took over the internet, brands turned to the newly found virtual space to promote themselves. This is a stepping stone into what is to come next as it taps into our day-to-day lives, economy, and social interactions [16]. The main component of web3 is the ability to build on it and interact with the web like never before and securely allowing users to maintain their privacy. Users are able to buy, maintain and sell virtual assets on the blockchain which unleashes a whole new era of using digital goods paved by the improvement of Virtual and Augmented realities known as VR and AR [1]. This has been proven by the famously known “Metaverse” world where companies such as Nintendo, Sony, and other gaming brands as well as celebrities such as Snoop Dogg rushed to create their own virtual theme parks and sold “plot of lands” in the said universe on the blockchain known as NFTs (Non-Fungible Tokens). This has created a ripple effect that continued in other virtually made assets, mainly video games, and lastly “The Metaverse” that “Meta” formerly known as “Facebook” is building upon.","PeriodicalId":371668,"journal":{"name":"2022 Ninth International Conference on Social Networks Analysis, Management and Security (SNAMS)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2022-11-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131634993","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 : 2022-11-29DOI: 10.1109/SNAMS58071.2022.10062698
Jaishree Ranganathan, Gloria Abuka
The increased availability of online feedback or review tools, and the enormous amount of information on these platforms, have made text summarization a vital research area in natural language processing. Instead of potential consumers going through thousands of reviews to get needed information, summarization will enable them to see a concise form of a chunk of reviews with relevant information. News and scientific articles have been used in text summarization models. This study proposes a text summarization method based on the Text-to- Text Transfer Transformer (T5) model. We use the University of California, Irvine's (UCI) drug reviews dataset. We manually created human summaries for the ten most useful reviews of a particular drug for 500 different drugs from the dataset. We fine-tune the Text-to- Text Transfer Transformer (T5) model to perform abstractive text summarization. The model's effectiveness was evaluated using the ROUGE metrics, and our model achieved an average of ROUGE1, ROUGE2, and ROUGEL scores of 45.62, 25.58, and 36.53, respectively. We also fine-tuned this model on a standard dataset(BBC News Dataset) previously used for text summarization and got average ROUGE1, ROUGE2, and ROUGEL scores of 69.05, 59.70, and 52.97, respectively.
{"title":"Text Summarization using Transformer Model","authors":"Jaishree Ranganathan, Gloria Abuka","doi":"10.1109/SNAMS58071.2022.10062698","DOIUrl":"https://doi.org/10.1109/SNAMS58071.2022.10062698","url":null,"abstract":"The increased availability of online feedback or review tools, and the enormous amount of information on these platforms, have made text summarization a vital research area in natural language processing. Instead of potential consumers going through thousands of reviews to get needed information, summarization will enable them to see a concise form of a chunk of reviews with relevant information. News and scientific articles have been used in text summarization models. This study proposes a text summarization method based on the Text-to- Text Transfer Transformer (T5) model. We use the University of California, Irvine's (UCI) drug reviews dataset. We manually created human summaries for the ten most useful reviews of a particular drug for 500 different drugs from the dataset. We fine-tune the Text-to- Text Transfer Transformer (T5) model to perform abstractive text summarization. The model's effectiveness was evaluated using the ROUGE metrics, and our model achieved an average of ROUGE1, ROUGE2, and ROUGEL scores of 45.62, 25.58, and 36.53, respectively. We also fine-tuned this model on a standard dataset(BBC News Dataset) previously used for text summarization and got average ROUGE1, ROUGE2, and ROUGEL scores of 69.05, 59.70, and 52.97, respectively.","PeriodicalId":371668,"journal":{"name":"2022 Ninth International Conference on Social Networks Analysis, Management and Security (SNAMS)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2022-11-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"117322169","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 : 2022-09-08DOI: 10.1109/SNAMS58071.2022.10062774
Li Miao, Jian Wu, Piyush Behre, Shuangyu Chang, S. Parthasarathy
It is challenging to train and deploy Transformer Language Models (LMs) for hybrid speech recognition second pass re-ranking in low-resource languages due to (1) data scarcity in low-resource languages, (2) expensive computing costs for training and refreshing 100+ monolingual models, and (3) hosting inefficiency considering sparse traffic. In this study, we present a novel way to group multiple low-resource locales together and optimize the performance of Multilingual Transformer LMs in ASR. Our Locale-group Multilingual Transformer LMs outperform traditional multilingual LMs along with reducing maintenance costs and operating expenses. Further, for high-traffic locales where deploying monolingual models is feasible, we show that fine-tuning our locale-group multilingual LMs produces better monolingual LM candidates than baseline monolingual LMs.
{"title":"Multilingual Transformer Language Model for Speech Recognition in Low-resource Languages","authors":"Li Miao, Jian Wu, Piyush Behre, Shuangyu Chang, S. Parthasarathy","doi":"10.1109/SNAMS58071.2022.10062774","DOIUrl":"https://doi.org/10.1109/SNAMS58071.2022.10062774","url":null,"abstract":"It is challenging to train and deploy Transformer Language Models (LMs) for hybrid speech recognition second pass re-ranking in low-resource languages due to (1) data scarcity in low-resource languages, (2) expensive computing costs for training and refreshing 100+ monolingual models, and (3) hosting inefficiency considering sparse traffic. In this study, we present a novel way to group multiple low-resource locales together and optimize the performance of Multilingual Transformer LMs in ASR. Our Locale-group Multilingual Transformer LMs outperform traditional multilingual LMs along with reducing maintenance costs and operating expenses. Further, for high-traffic locales where deploying monolingual models is feasible, we show that fine-tuning our locale-group multilingual LMs produces better monolingual LM candidates than baseline monolingual LMs.","PeriodicalId":371668,"journal":{"name":"2022 Ninth International Conference on Social Networks Analysis, Management and Security (SNAMS)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2022-09-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131981167","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：