With the widespread adoption of multimedia web APIs (API) in web and mobile applications, a substantial proliferation of these APIs is observed. These APIs have streamlined development processes, reducing both time and costs. Nevertheless, identifying the required APIs from the vast array of options has emerged as a significant challenge. Collaborative filtering (CF)-based recommendation technologies have demonstrated their efficiency in presenting developers with potentially useful APIs. However, these methods often suffer from popularity bias, i.e., popular APIs tend to dominate the recommendation lists. This imbalance in recommendation opportunities among APIs hinders the growth of the multimedia API ecosystem. To mitigate the popularity bias produced by CF-based API recommendation methods, this article introduces a novel debiasing strategy that combines a log postprocessing adjustment (LPA) with determinant point process (DPP). Specifically, the LPA is employed during the prediction phase to yield a more balanced set of candidate APIs. Then, DPP is utilized to generate recommendation lists that are not just relevant but also diverse in terms of API popularity. Experimental results reveal that our proposed method surpasses existing state-of-the-art approaches in multimedia API recommendation, excelling in both accuracy and the capability to mitigate popularity bias effectively.
{"title":"Counteracting Popularity Bias in Multimedia Web API Recommendation","authors":"Dengshuai Zhai;Chao Yan;Weiyi Zhong;Shaoqi Ding;Lianyong Qi;Xiaokang Zhou","doi":"10.1109/TCSS.2024.3517601","DOIUrl":"https://doi.org/10.1109/TCSS.2024.3517601","url":null,"abstract":"With the widespread adoption of multimedia web APIs (API) in web and mobile applications, a substantial proliferation of these APIs is observed. These APIs have streamlined development processes, reducing both time and costs. Nevertheless, identifying the required APIs from the vast array of options has emerged as a significant challenge. Collaborative filtering (CF)-based recommendation technologies have demonstrated their efficiency in presenting developers with potentially useful APIs. However, these methods often suffer from popularity bias, i.e., popular APIs tend to dominate the recommendation lists. This imbalance in recommendation opportunities among APIs hinders the growth of the multimedia API ecosystem. To mitigate the popularity bias produced by CF-based API recommendation methods, this article introduces a novel debiasing strategy that combines a log postprocessing adjustment (LPA) with determinant point process (DPP). Specifically, the LPA is employed during the prediction phase to yield a more balanced set of candidate APIs. Then, DPP is utilized to generate recommendation lists that are not just relevant but also diverse in terms of API popularity. Experimental results reveal that our proposed method surpasses existing state-of-the-art approaches in multimedia API recommendation, excelling in both accuracy and the capability to mitigate popularity bias effectively.","PeriodicalId":13044,"journal":{"name":"IEEE Transactions on Computational Social Systems","volume":"12 5","pages":"3858-3868"},"PeriodicalIF":4.5,"publicationDate":"2025-01-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145230020","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-01-28DOI: 10.1109/TCSS.2025.3531587
{"title":"IEEE Transactions on Computational Social Systems Publication Information","authors":"","doi":"10.1109/TCSS.2025.3531587","DOIUrl":"https://doi.org/10.1109/TCSS.2025.3531587","url":null,"abstract":"","PeriodicalId":13044,"journal":{"name":"IEEE Transactions on Computational Social Systems","volume":"12 1","pages":"C2-C2"},"PeriodicalIF":4.5,"publicationDate":"2025-01-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=10856570","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143361026","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-01-28DOI: 10.1109/TCSS.2025.3531589
{"title":"IEEE Systems, Man, and Cybernetics Society Information","authors":"","doi":"10.1109/TCSS.2025.3531589","DOIUrl":"https://doi.org/10.1109/TCSS.2025.3531589","url":null,"abstract":"","PeriodicalId":13044,"journal":{"name":"IEEE Transactions on Computational Social Systems","volume":"12 1","pages":"C3-C3"},"PeriodicalIF":4.5,"publicationDate":"2025-01-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=10856572","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143106470","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-01-28DOI: 10.1109/TCSS.2025.3531591
{"title":"IEEE Transactions on Computational Social Systems Information for Authors","authors":"","doi":"10.1109/TCSS.2025.3531591","DOIUrl":"https://doi.org/10.1109/TCSS.2025.3531591","url":null,"abstract":"","PeriodicalId":13044,"journal":{"name":"IEEE Transactions on Computational Social Systems","volume":"12 1","pages":"C4-C4"},"PeriodicalIF":4.5,"publicationDate":"2025-01-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=10856569","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143106471","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-01-28DOI: 10.1109/TCSS.2025.3530618
Ziwen Sun;Jian Kang;Kun Qian;Björn W. Schuller;Bin Hu
{"title":"Creating Healthier Living Environments: The Role of Soundscapes in Promoting Mental Health and Well-Being","authors":"Ziwen Sun;Jian Kang;Kun Qian;Björn W. Schuller;Bin Hu","doi":"10.1109/TCSS.2025.3530618","DOIUrl":"https://doi.org/10.1109/TCSS.2025.3530618","url":null,"abstract":"","PeriodicalId":13044,"journal":{"name":"IEEE Transactions on Computational Social Systems","volume":"12 1","pages":"2-10"},"PeriodicalIF":4.5,"publicationDate":"2025-01-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=10856573","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143360871","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-01-27DOI: 10.1109/TCSS.2024.3504357
Vaibhav Garg;Ganning Xu;Munindar P. Singh
Inciting speech seeks to instill hostility or anger in readers or motivate them to take action against a target group. Whereas hate speech in social media has garnered much attention, inciting speech has not been well studied in domains such as religion. We address two aspects of religious incitement: 1) what rhetorical strategies are used in it?; and 2) do the same strategies apply across disparate social contexts and targets? We identify inciting speech against Muslims but demonstrate the generality of the construct vis à vis other targets. We adopt existing datasets of Islamophobic WhatsApp posts and hateful and offensive posts (Twitter and Gab) against other targets. Our methods include: 1) qualitative analysis revealing rhetorical strategies; and 2) an iterative process to label the data, yielding a tool to detect incitement. Incitement applies three rhetorical strategies focused, respectively, on the target group's identity, their imputed misdeeds, and an exhortation to act against them. These strategies carry distinct textual signatures. Our tool (with additional verification) reveals that inciting sentences appear in non-Islamophobic posts and in other contexts (e.g., posts against certain gender identities), indicating the generality of incitement as a concept. Incitement reflects a wide swath of malicious speech omitted from traditional analyses. Understanding and identifying incitement can facilitate online moderation and thus concomitantly reduce harm in real life.
{"title":"Understanding Inciting Speech as New Malice","authors":"Vaibhav Garg;Ganning Xu;Munindar P. Singh","doi":"10.1109/TCSS.2024.3504357","DOIUrl":"https://doi.org/10.1109/TCSS.2024.3504357","url":null,"abstract":"Inciting speech seeks to instill hostility or anger in readers or motivate them to take action against a target group. Whereas hate speech in social media has garnered much attention, inciting speech has not been well studied in domains such as religion. We address two aspects of religious incitement: 1) what rhetorical strategies are used in it?; and 2) do the same strategies apply across disparate social contexts and targets? We identify inciting speech against Muslims but demonstrate the generality of the construct vis à vis other targets. We adopt existing datasets of Islamophobic WhatsApp posts and hateful and offensive posts (Twitter and Gab) against other targets. Our methods include: 1) qualitative analysis revealing rhetorical strategies; and 2) an iterative process to label the data, yielding a tool to detect incitement. Incitement applies three rhetorical strategies focused, respectively, on the target group's identity, their imputed misdeeds, and an exhortation to act against them. These strategies carry distinct textual signatures. Our tool (with additional verification) reveals that inciting sentences appear in non-Islamophobic posts and in other contexts (e.g., posts against certain gender identities), indicating the generality of incitement as a concept. Incitement reflects a wide swath of malicious speech omitted from traditional analyses. Understanding and identifying incitement can facilitate online moderation and thus concomitantly reduce harm in real life.","PeriodicalId":13044,"journal":{"name":"IEEE Transactions on Computational Social Systems","volume":"12 3","pages":"947-956"},"PeriodicalIF":4.5,"publicationDate":"2025-01-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144179157","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
This survey article explores graph-based approaches to multimodal human activity recognition in indoor environments, emphasizing their relevance to advancing multimodal representation and reasoning. With the growing importance of integrating diverse data sources such as sensor events, contextual information, and spatial data, effective human activity recognition methods are essential for applications in smart homes, digital health, and more. We review various graph-based techniques, highlighting their strengths in encoding complex relationships and improving activity recognition performance. Furthermore, we discuss the computational efficiencies and generalization capabilities of these methods across different environments. By providing a comprehensive overview of the state-of-the-art in graph-based human activity recognition, this article aims to contribute to the development of more accurate, interpretable, and robust multimodal systems for understanding human activities in indoor settings.
{"title":"Graph-Based Methods for Multimodal Indoor Activity Recognition: A Comprehensive Survey","authors":"Saeedeh Javadi;Daniele Riboni;Luigi Borzì;Samaneh Zolfaghari","doi":"10.1109/TCSS.2024.3523240","DOIUrl":"https://doi.org/10.1109/TCSS.2024.3523240","url":null,"abstract":"This survey article explores graph-based approaches to multimodal human activity recognition in indoor environments, emphasizing their relevance to advancing multimodal representation and reasoning. With the growing importance of integrating diverse data sources such as sensor events, contextual information, and spatial data, effective human activity recognition methods are essential for applications in smart homes, digital health, and more. We review various graph-based techniques, highlighting their strengths in encoding complex relationships and improving activity recognition performance. Furthermore, we discuss the computational efficiencies and generalization capabilities of these methods across different environments. By providing a comprehensive overview of the state-of-the-art in graph-based human activity recognition, this article aims to contribute to the development of more accurate, interpretable, and robust multimodal systems for understanding human activities in indoor settings.","PeriodicalId":13044,"journal":{"name":"IEEE Transactions on Computational Social Systems","volume":"12 5","pages":"3728-3746"},"PeriodicalIF":4.5,"publicationDate":"2025-01-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=10843961","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145255997","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-01-13DOI: 10.1109/TCSS.2024.3506158
Zekun Yang;Kun Lv;Jian Shu;Zheng Li;Ping Xiao
In recent years, large language models (LLMs) represented by GPT-4 have achieved tremendous success in natural language-centered tasks. Nevertheless, LLMs face inherent challenges in tasks involving both natural language and molecular modalities. Although there has been some research progress on these tasks, two challenges remain unresolved: modeling the differences in representation format between natural language and molecular modalities, and capturing subtle differences under an instruction-tuned paradigm. To address these two challenges, this article proposes a two-stage training framework to build molecular knowledge-enhanced LLM, named Mol-LLM. The first stage utilizes the multitask instruction tuning method to tackle the modality differences between natural language and molecular sequence. The second stage is the direct preference optimization training strategy with three random preference actions to capture subtle differences under the instruction-tuned paradigm. Extensive experiments have demonstrated the state-of-the-art performances of the model Mol-LLM proposed in this study, including mol2mol, text2text, mol2text, and text2mol task. The effectiveness of the module has been further verified through ablation studies, and the generalizability has been confirmed by additional supportive experiments.
{"title":"Incorporating Molecular Knowledge in Large Language Models via Multimodal Modeling","authors":"Zekun Yang;Kun Lv;Jian Shu;Zheng Li;Ping Xiao","doi":"10.1109/TCSS.2024.3506158","DOIUrl":"https://doi.org/10.1109/TCSS.2024.3506158","url":null,"abstract":"In recent years, large language models (LLMs) represented by GPT-4 have achieved tremendous success in natural language-centered tasks. Nevertheless, LLMs face inherent challenges in tasks involving both natural language and molecular modalities. Although there has been some research progress on these tasks, two challenges remain unresolved: modeling the differences in representation format between natural language and molecular modalities, and capturing subtle differences under an instruction-tuned paradigm. To address these two challenges, this article proposes a two-stage training framework to build molecular knowledge-enhanced LLM, named Mol-LLM. The first stage utilizes the multitask instruction tuning method to tackle the modality differences between natural language and molecular sequence. The second stage is the direct preference optimization training strategy with three random preference actions to capture subtle differences under the instruction-tuned paradigm. Extensive experiments have demonstrated the state-of-the-art performances of the model Mol-LLM proposed in this study, including <italic>mol2mol</i>, <italic>text2text</i>, <italic>mol2text</i>, and <italic>text2mol</i> task. The effectiveness of the module has been further verified through ablation studies, and the generalizability has been confirmed by additional supportive experiments.","PeriodicalId":13044,"journal":{"name":"IEEE Transactions on Computational Social Systems","volume":"12 5","pages":"3660-3670"},"PeriodicalIF":4.5,"publicationDate":"2025-01-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145230027","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-01-13DOI: 10.1109/TCSS.2024.3524297
Suping Wang;Fei Zhou;Ming Yang;Lei Shi;Chaohong Tan
Cross-modal retrieval is crucial for achieving accurate and efficient information retrieval by establishing semantic correlations between heterogeneous images and text. However, traditional image-text training sets suffer from information asymmetry, which includes short lengths and limited sentence structures. This phenomenon often results in insufficient representations of essential visual information. We introduce RichDataset, which offers extensive semantic information. It includes diverse real-life image-text pairs and AI-generated content across domains such as news, entertainment, education, and posters. Compared with classic benchmarks such as Flickr30k and MS-COCO, RichDataset exhibits a novel and balanced distribution. Existing cross-modal retrieval models face challenges in extracting distinct features from the emerging data, leading to low retrieval accuracy. We propose SGG-MVAR, a comprehensive retrieval model guided by multiview scene information and semantic relationships. Leveraging a scene knowledge database, our model parses scene graphs and identifies differences in attributes and relationships. We conduct extensive experiments to evaluate our proposed dataset and model. All experimental results consistently demonstrate a significant improvement in recall for cross-modal retrieval.
{"title":"SGG-MVAR: Cross-Modal Retrieval With Scene Graph Generation and Multiview Attribute Relationship Guidance","authors":"Suping Wang;Fei Zhou;Ming Yang;Lei Shi;Chaohong Tan","doi":"10.1109/TCSS.2024.3524297","DOIUrl":"https://doi.org/10.1109/TCSS.2024.3524297","url":null,"abstract":"Cross-modal retrieval is crucial for achieving accurate and efficient information retrieval by establishing semantic correlations between heterogeneous images and text. However, traditional image-text training sets suffer from information asymmetry, which includes short lengths and limited sentence structures. This phenomenon often results in insufficient representations of essential visual information. We introduce RichDataset, which offers extensive semantic information. It includes diverse real-life image-text pairs and AI-generated content across domains such as news, entertainment, education, and posters. Compared with classic benchmarks such as Flickr30k and MS-COCO, RichDataset exhibits a novel and balanced distribution. Existing cross-modal retrieval models face challenges in extracting distinct features from the emerging data, leading to low retrieval accuracy. We propose SGG-MVAR, a comprehensive retrieval model guided by multiview scene information and semantic relationships. Leveraging a scene knowledge database, our model parses scene graphs and identifies differences in attributes and relationships. We conduct extensive experiments to evaluate our proposed dataset and model. All experimental results consistently demonstrate a significant improvement in recall for cross-modal retrieval.","PeriodicalId":13044,"journal":{"name":"IEEE Transactions on Computational Social Systems","volume":"12 5","pages":"3671-3683"},"PeriodicalIF":4.5,"publicationDate":"2025-01-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145230079","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-01-13DOI: 10.1109/TCSS.2024.3455415
Marco Furini;Luca Mariotti;Riccardo Martoglia;Manuela Montangero
This study explores the influence of social media on health-related discourse amid the COVID-19 pandemic, focusing on Italian-language tweets posted on Twitter from March 2020 to December 2021. Analyzing a dataset comprising 13 million tweets, the research addresses three key questions: who emerged as opinion leaders on Twitter during the pandemic in Italy?; did health institutions in Italy successfully establish themselves as opinion leaders?; and how did the content of COVID-19-related tweets in Italy evolve over time? Employing a custom-designed graph and the personalized PageRank algorithm, the study identifies opinion leaders on Twitter. Additionally, psycholinguistic analysis provides insights into the content, themes, and emotional undertones of the tweets. The findings of this research contribute to a deeper understanding of social media's influence on public opinion and behavior during the pandemic. Furthermore, they offer valuable insights for public health officials and policymakers seeking to address health-related issues on social media platforms.
{"title":"A Novel Graph-Based Approach to Identify Opinion Leaders in Twitter","authors":"Marco Furini;Luca Mariotti;Riccardo Martoglia;Manuela Montangero","doi":"10.1109/TCSS.2024.3455415","DOIUrl":"https://doi.org/10.1109/TCSS.2024.3455415","url":null,"abstract":"This study explores the influence of social media on health-related discourse amid the COVID-19 pandemic, focusing on Italian-language tweets posted on Twitter from March 2020 to December 2021. Analyzing a dataset comprising 13 million tweets, the research addresses three key questions: who emerged as opinion leaders on Twitter during the pandemic in Italy?; did health institutions in Italy successfully establish themselves as opinion leaders?; and how did the content of COVID-19-related tweets in Italy evolve over time? Employing a custom-designed graph and the personalized PageRank algorithm, the study identifies opinion leaders on Twitter. Additionally, psycholinguistic analysis provides insights into the content, themes, and emotional undertones of the tweets. The findings of this research contribute to a deeper understanding of social media's influence on public opinion and behavior during the pandemic. Furthermore, they offer valuable insights for public health officials and policymakers seeking to address health-related issues on social media platforms.","PeriodicalId":13044,"journal":{"name":"IEEE Transactions on Computational Social Systems","volume":"12 3","pages":"1268-1278"},"PeriodicalIF":4.5,"publicationDate":"2025-01-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144186007","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","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：