2022 International Conference on Technologies and Applications of Artificial Intelligence (TAAI)最新文献

{"title":"Visualization of the Relationship Between Lectures and Laboratories Using SSNMF","authors":"Kyoka Yamamoto, Ryosuke Yamanishi, Mitsunori Matsushita","doi":"10.1109/TAAI57707.2022.00036","DOIUrl":"https://doi.org/10.1109/TAAI57707.2022.00036","url":null,"abstract":"This study aims to visualize the relationship between lectures and fields of specialization (laboratories) so that students can choose lectures with a future direction. Since the university curriculum is highly flexible; students choose their own lectures. Taking into account their own objectives, students select the basic knowledge necessary for their purposes. However, it is difficult for students without sufficient knowledge to understand their relevance from the syllabus. The purpose of this study is to propose a method for estimating the relevance between lectures and laboratories in an undergraduate school as a help to provide students with an objective analysis of lectures, i.e., not only knowledge but also examples of its use. The proposed method applies a semi-supervised non-negative matrix factorization to identify common factors of knowledge in each combination of lecture and laboratory. It is suggested that the proposed method calculates reasonable results for the relationship between lectures and laboratories.","PeriodicalId":111620,"journal":{"name":"2022 International Conference on Technologies and Applications of Artificial Intelligence (TAAI)","volume":"52 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115734782","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}

{"title":"Generate Multi-Perspective Summarization with Pairwise Alignment Mechanism","authors":"Ching-Shiuan Hsiao, Jiun-An Tsai, Jung-Shiuan Liou, Yi-Shin Chen","doi":"10.1109/TAAI57707.2022.00024","DOIUrl":"https://doi.org/10.1109/TAAI57707.2022.00024","url":null,"abstract":"Cyberterrorism is often defined as attack using communication networks to cause destruction or generate fear in society to achieve political or ideological goals. Nowadays, due to the accessibility and anonymity, social media has become the breeding grounds for spreading cyberterrorism. To counter cyberterrorism efficiently and effectively, it is critical for the legal enforcement to summarize information on the internet. Particularly, to analyze intelligence comprehensively and make decision without bias, there are multiple perspectives to be considered. However, it remains two challenges to be solved for multi-perspective summarization, which are lacking of annotations and various perspectives with conflicts. To address the challenges, we propose Pairwise Alignment Mechanism and Perspective Consensus Strategy to incorporate different perspectives into the summary. The experiment results show that our methodology can significantly leverage the coverage of multiple perspectives and maintain the consistency of the generated summary.","PeriodicalId":111620,"journal":{"name":"2022 International Conference on Technologies and Applications of Artificial Intelligence (TAAI)","volume":"4 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127551733","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}

{"title":"Two-Phase Optimization for Shift Scheduling in Individualized Teaching Cram School","authors":"Yosuke Suzuki, Kazunori Mizuno","doi":"10.1109/TAAI57707.2022.00031","DOIUrl":"https://doi.org/10.1109/TAAI57707.2022.00031","url":null,"abstract":"A shift schedule is indispensable for managing the work of each employee. However, it is difficult to create a shift schedule manually that takes into account employee preferences. Individualized teaching cram schools also requires the construction of a shift schedule for teachers, but at the very least, the subjects that students take and the subjects that teachers are able to teach. It is thus necessary to construct a schedule that satisfies both requests of teachers and students. In this paper, we propose a two-phase optimization method to find such a schedule: shift scheduling using genetic algorithms and student timetabling using simulated annealing. We demonstrate that our proposed method can construct a valid and feasible schedule for each of teachers and students using an actual dataset.","PeriodicalId":111620,"journal":{"name":"2022 International Conference on Technologies and Applications of Artificial Intelligence (TAAI)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130028274","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}

{"title":"Decoupled Knowledge Propagation Network and Sampling Strategies for Recommendation System","authors":"Zhen-Yuan Kuo, Bor-Shen Lin","doi":"10.1109/TAAI57707.2022.00037","DOIUrl":"https://doi.org/10.1109/TAAI57707.2022.00037","url":null,"abstract":"State-of-the-art knowledge-graph-based models, such as RippleNet and CKAN, have been successfully applied to recommendation systems. These models make use of knowledge graph to expand the entity information, which is similar to ripples, so as to strengthen the correlation between the user's preference and the candidate item. In RippleNet and CKAN, however, the networks for learning the features of user preference and candidate item are mutually coupled, which might limit the recommendation performance, especially for multi-levels propagation. On the other hand, when entities are expanded over the knowledge graph, the amount of entities increases exponentially with the number of diffusion levels. In earlier research, random sampling was adopted to restrict the number of triples, which however may limit the coverage of the expanded knowledge for model learning as well as the prediction performance. To tackle these problems, we propose a decoupled knowledge propagation network that separates the user's preference and the candidate item to distinguish the candidate items better. Experiments conducted on the MovieLens 1M recommendation dataset and the Microsoft Satori knowledge graph show the proposed mode outperforms RippleNet and CKAN. In addition, three sampling strategies, including balanced sampling, non-duplication, and dynamic sampling were proposed to deal with the sampling issue, and experiments show such sampling strategies are effective and superimposed for the prediction performance.","PeriodicalId":111620,"journal":{"name":"2022 International Conference on Technologies and Applications of Artificial Intelligence (TAAI)","volume":"119 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124586529","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}

{"title":"Combining Conditional Generative Adversarial Networks and YOLOv4 for Mango Classification","authors":"Li-Hua Li, Ling-Qi Jiang, Yu-Fang Peng, Ye-Shan Liu, Kai-Lun Chung","doi":"10.1109/TAAI57707.2022.00019","DOIUrl":"https://doi.org/10.1109/TAAI57707.2022.00019","url":null,"abstract":"Mango is one of the most important exporting products in Taiwan and it has provided Taiwan with numerous economic benefits. To maximize the value, mangos are usually classified into grade A, grade B, and grade C before they are sent to exporting process, domestic sale, or fruit punching factory, respectively. Since the freshness of mangos is limited by time, it is important to accelerate the classification process. To improve the classification performance, this research proposes the Irwin mango classification system (IMCS) by combining Conditional Generative Adversarial Network (CGAN) and YOLOv4. CGAN is applied to help data expansion when the amount of data is insufficient. YOLOv4 is applied to classify the grade of mango and to detect the location of mango. To show our approach is better than other methods, we compare our model with YOLOv3 SPP, YOLOv4, ResNet, AlexNet, VGG 16, DenseNet, ShuffleNet, Fusion model, and the best outcome in AI CUP. The experiments show that our CGAN plus YOLOv4 (CGAN+YOLOv4) model has the best outcome with 85% precision for training, 82% precision for testing, and 90.07% for mAP. The weighted average recall (WAR) is 84.83% which is better than the best outcome of AI CUP 2020.","PeriodicalId":111620,"journal":{"name":"2022 International Conference on Technologies and Applications of Artificial Intelligence (TAAI)","volume":"43 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131313432","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}

{"title":"Federated Learning Operations (FLOps): Challenges, Lifecycle and Approaches","authors":"Qi Cheng, Guodong Long","doi":"10.1109/TAAI57707.2022.00012","DOIUrl":"https://doi.org/10.1109/TAAI57707.2022.00012","url":null,"abstract":"Federated Learning has witnessed a rapid growth in research and industry applications as it offers the benefits of privacy preserving while contributing to the global model training. Cross-silo federated learning systems which are usually geographically distributed and cross-organizational are becoming a reality. Although DevOps and MLOps methodologies may help improving traditional machine learning systems' development efficiency and productivity, it is still challenging for them to develop cross-silo federated learning systems in a productive way. In this paper, we propose FLOps (Federated Learning Operations), a new methodology for developing cross-silo federated learning systems efficiently and continuously. By elaborating the challenges that FLOps is facing, we construct the lifecycle of FLOps, and propose approaches to FLOps. Finally, we highlight potential research directions of FLOps.","PeriodicalId":111620,"journal":{"name":"2022 International Conference on Technologies and Applications of Artificial Intelligence (TAAI)","volume":"52 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122235879","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}

{"title":"Proposal of Consistent Learning Model with Exchange Monte Carlo","authors":"H. Shibata, Y. Takama","doi":"10.1109/TAAI57707.2022.00044","DOIUrl":"https://doi.org/10.1109/TAAI57707.2022.00044","url":null,"abstract":"This paper proposes a consistent learning model based on Exchange Monte Carlo Method. The paper also gives discussion with respect to experiments on the synthesized case. Learning model is currently focusing on the model with the interface for input and output. On that model, preparing a dataset remains in human's work, and there is still not sufficient research how to prepare a valuable dataset efficiently. Exchange Monte Carlo is used widely for both purposes of optimization and estimation of a probability distribution, and it has the ability to combine any probability model in one and sample the model's state efficiency from the combined probability model. From this point of view, when we consider the three models, i.e., real space, learnt model, and model of parameter distribution, we can combine them and construct the consistent model that explains the phenomena of learning consistently. It is supposed that those samples give us valuable dataset and set of the learnt parameter, when the parameter space also modeled with Bayesian inference framework. With these ideas mentioned so far, this paper proposes a generalized consistent probability model of the real space, learning model, and parameters' distribution of the learning model. To challenge to the sampling problem on high-dimensionality of the consistent model, Exchange Monte Carlo and Hamiltonian dynamics are employed. Experiments show the proposed method works on the synthesized case, that is, the original distribution is approximated well and parameter is optimized too, only by sampling from the consistent model without preparing dataset.","PeriodicalId":111620,"journal":{"name":"2022 International Conference on Technologies and Applications of Artificial Intelligence (TAAI)","volume":"131 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114827494","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}

{"title":"Deep Recognition of Facial Expressions in Movies","authors":"Lieu-Hen Chen, Eric Hsiao-Kuang Wu, Eri Shimokawara, Hao-Ming Hung, Wei-Chek Ong-Lim","doi":"10.1109/TAAI57707.2022.00020","DOIUrl":"https://doi.org/10.1109/TAAI57707.2022.00020","url":null,"abstract":"Consumer feedback is used in lots of fields for various purposes. However, traditional paper questionnaires or online surveys cannot fully meet these demands for obtaining accurate and useful feedback from consumers. Therefore, we propose a deep learning based deep recognition of facial micro expressions in order to receive more realistic feedback from users in this study. To achieve this goal, we integrate several approaches including: 1. using trained face detection model to capture face image from input. 2. training a high accurate 468-point landmark detection model with multiple face dataset. Based on the FACS (Facial Action Coding System) table, we categorize these landmarks into 13 groups of facial regions. These regions with specific emotion labels are used as our target units of AU (Action Unit) detection. 3. training CNN model to detect and analysis AUs from facial landmark data. 4. implying FACS to evaluate the facial expressions and emotions. and 5. using a straightforward GUI plotter to show the digitized emotions. The experiment results show that not only the primary emotion but also the secondary emotion of users in movies can be detected and evaluated successfully. Therefore, our system has a great potential for detecting micro expressions in a more accurate and comprehensive manner.","PeriodicalId":111620,"journal":{"name":"2022 International Conference on Technologies and Applications of Artificial Intelligence (TAAI)","volume":"38 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129046325","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}

{"title":"Stress-Coping Tweets Acquisition: A Two-phase Bootstrapping Method on Patterns and Semantic Features","authors":"Jui-Ching Weng, Yen-Hao Huang, Kezia Flaviana Irene Tamus, Y. Lien, Yi-Shin Chen","doi":"10.1109/TAAI57707.2022.00029","DOIUrl":"https://doi.org/10.1109/TAAI57707.2022.00029","url":null,"abstract":"Stress is integral to biological survival. However, without an appropriate coping response, high stress levels and long-term stressful situations may lead to negative mental health outcomes. Since the COVID-19 pandemic, remote assessment of mental health has become imperative. The majority of past studies focused on detecting users' stress levels rather than coping responses using social media. Because of the diversity of human expression and because people do not usually express stress and the corresponding coping response simultaneously, it is challenging to extract users' tweets about their coping responses to stressful events from their daily tweets. Consequently, there are two goals being pursued in this study: to anchor users' stress statuses and to detect their stress responses based on the existing stressful conditions. In order to accomplish these goals, we propose a framework that consists of two phases: the construction of stress dataset and the extraction of coping responses. Since the stressed users' data are lacking, the first phase is to construct a stress dataset based on stress-related hashtags, personal pronouns, and emotion recognition. In addition, to ensure the collection of enough tweets to observe the coping responses of stressed users, we broadened the survey's scope by collecting all tweets from the same user. In the second phase, stress-coping tweets were extracted by utilizing bootstrapping-based patterns and semantic features. The bootstrapping method was used to enrich word patterns for text expression and the semantic feature to assess the meaning of sentences. The collected data included the tweets of the stressed users identified in Phase 1 and the various coping responses from Phase 2 can contribute to developing a tool for the remote assessment of mental health. The experimental results show that our two-phase method outperforms the baseline and can help improve the efficiency of extracting stress-coping tweets.","PeriodicalId":111620,"journal":{"name":"2022 International Conference on Technologies and Applications of Artificial Intelligence (TAAI)","volume":"81 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124665505","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}

{"title":"Relation-Aware Image Captioning for Explainable Visual Question Answering","authors":"Ching-Shan Tseng, Ying-Jia Lin, Hung-Yu kao","doi":"10.1109/TAAI57707.2022.00035","DOIUrl":"https://doi.org/10.1109/TAAI57707.2022.00035","url":null,"abstract":"Recent studies leveraging object detection models for Visual Question Answering (VQA) ignore the correlations or interactions between multiple objects. In addition, the previous VQA models are black boxes for human beings, which means it is difficult to explain why a model returns correct or wrong answers. To solve the problems above, we propose a new model structure with image captioning for the VQA task. Our model constructs a relation graph according to the relative positions between region pairs and then produces relation-aware visual features with a relation encoder. To make the predictions explainable, we introduce an image captioning module and conduct a multi-task training process. In the meantime, the generated captions are injected into the predictor to assist cross-modal understanding. Experiments show that our model can generate meaningful answers and explanations according to the questions and images. Besides, the relation encoder and the caption-attended predictor lead to improvement for different types of questions.","PeriodicalId":111620,"journal":{"name":"2022 International Conference on Technologies and Applications of Artificial Intelligence (TAAI)","volume":"58 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121806465","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}