Tosh Yamamoto, A. Liao, W. Wu, Meilun Shih, Ju-Ling Shih, Hui-Chun Chu
This paper purports to share with the higher education community the global PBL active learning curriculum and the learning environment, which have been collaboratively developed with the universities in Taiwan and Kansai University (KU). The collaborated universities developed an optimal curriculum to enhance and nurture students' "Future Work Skills 2020" defined by the Institute for the Future, such future human skills as Sense Making, Social Intelligence, Novel & Adaptive Thinking, Cross-Cultural Competencies, Computational Thinking, New Media Literacy, Transdisciplinarity, Design Mindset, Cognitive Load Management, and Virtual Collaboration. The curriculum fully employs PBL strategies in global teams, where teams for PBL are organized with students with heterogeneous cultural backgrounds in the virtual learning environment. The basic concept of such curriculum is based on COIL (Collaborative Online International Learning), originally developed by State University of New York. COIL makes full usage of IT to generate virtual learning environment for students worldwide. In order to go beyond the COIL concept incorporating the future skills defined by IFTF, the allied universities employed PBL in global AGILE teams to deepen insights from various cultural viewpoints in terms of consensus building through team discussions. Due to the spatial and temporal differences, enrolled students conducted their team learning activities in the virtual learning environment asynchronously, making use of IT technologies and cloud services in order to be on "the same page" in the progress of the project throughout the course. Further, the assessment strategies to enhance students' efficacy is the key factor in the course, which is also discussed with examples. This paper reports the global PBL active learning curriculum and environment collaboratively developed with the universities in Taiwan and Kansai University.
{"title":"A Proposal for the Global and Collaborative PBL Learning Environment Where All Global Members on Different Campuses Are \"On the Same Page\" throughout the Process of Learning in the Project","authors":"Tosh Yamamoto, A. Liao, W. Wu, Meilun Shih, Ju-Ling Shih, Hui-Chun Chu","doi":"10.1109/TAAI.2018.00029","DOIUrl":"https://doi.org/10.1109/TAAI.2018.00029","url":null,"abstract":"This paper purports to share with the higher education community the global PBL active learning curriculum and the learning environment, which have been collaboratively developed with the universities in Taiwan and Kansai University (KU). The collaborated universities developed an optimal curriculum to enhance and nurture students' \"Future Work Skills 2020\" defined by the Institute for the Future, such future human skills as Sense Making, Social Intelligence, Novel & Adaptive Thinking, Cross-Cultural Competencies, Computational Thinking, New Media Literacy, Transdisciplinarity, Design Mindset, Cognitive Load Management, and Virtual Collaboration. The curriculum fully employs PBL strategies in global teams, where teams for PBL are organized with students with heterogeneous cultural backgrounds in the virtual learning environment. The basic concept of such curriculum is based on COIL (Collaborative Online International Learning), originally developed by State University of New York. COIL makes full usage of IT to generate virtual learning environment for students worldwide. In order to go beyond the COIL concept incorporating the future skills defined by IFTF, the allied universities employed PBL in global AGILE teams to deepen insights from various cultural viewpoints in terms of consensus building through team discussions. Due to the spatial and temporal differences, enrolled students conducted their team learning activities in the virtual learning environment asynchronously, making use of IT technologies and cloud services in order to be on \"the same page\" in the progress of the project throughout the course. Further, the assessment strategies to enhance students' efficacy is the key factor in the course, which is also discussed with examples. This paper reports the global PBL active learning curriculum and environment collaboratively developed with the universities in Taiwan and Kansai University.","PeriodicalId":211734,"journal":{"name":"2018 Conference on Technologies and Applications of Artificial Intelligence (TAAI)","volume":"98 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116459414","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}
In recent years, we have witnessed a sudden increase in mosquito-borne diseases and related casualties. This makes it important to have an efficient mosquito classification system. In this paper, we implement a mosquito classification system which is capable of identifying Aedes and Culex (types of the mosquito) automatically. To facilitate the implementation of such Internet of Things (IoT) based system, we first create a trap device with a stable area for filming mosquitoes. Then, we analyze video frames in order to reduce the video size for transmission. We also build a model to identify different types of mosquitoes using deep learning. Later, we fine-tune the edge computing on the trap device to optimize the system efficiency. Finally, we integrate the device and the model into a mosquito classification system and test the system in wild fields in Taiwan. The tests show significant results when the experiments are conducted in the rural area. We are able to achieve an accuracy of 98% for validation data and 90.5% for testing data.
{"title":"A Vector Mosquitoes Classification System Based on Edge Computing and Deep Learning","authors":"Li-Pang Huang, Ming-Hong Hong, Cyuan-Heng Luo, Sachit Mahajan, Ling-Jyh Chen","doi":"10.1109/TAAI.2018.00015","DOIUrl":"https://doi.org/10.1109/TAAI.2018.00015","url":null,"abstract":"In recent years, we have witnessed a sudden increase in mosquito-borne diseases and related casualties. This makes it important to have an efficient mosquito classification system. In this paper, we implement a mosquito classification system which is capable of identifying Aedes and Culex (types of the mosquito) automatically. To facilitate the implementation of such Internet of Things (IoT) based system, we first create a trap device with a stable area for filming mosquitoes. Then, we analyze video frames in order to reduce the video size for transmission. We also build a model to identify different types of mosquitoes using deep learning. Later, we fine-tune the edge computing on the trap device to optimize the system efficiency. Finally, we integrate the device and the model into a mosquito classification system and test the system in wild fields in Taiwan. The tests show significant results when the experiments are conducted in the rural area. We are able to achieve an accuracy of 98% for validation data and 90.5% for testing data.","PeriodicalId":211734,"journal":{"name":"2018 Conference on Technologies and Applications of Artificial Intelligence (TAAI)","volume":"161 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115996337","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}
Monte-Carlo tree search (MCTS) algorithms play an important role in developing computer players for many games. The performance of MCTS players is often leveraged in combination with offline knowledge, i.e., evaluation functions. In particular, recently AlphaGo and AlphaGo Zero achieved a big success in developing strong computer Go player by combining evaluation functions consisting of deep neural networks with a variant of PUCT (Predictor + UCB applied to trees). The effect of evaluation functions on the strength of MCTS algorithms, however, has not been investigated well, especially in terms of the quality of evaluation functions. In this study, we address this issue and empirically analyze the AlphaGo's PUCT algorithm by using Othello (Reversi) as the target game. We investigate the strength of PUCT players using variants of an existing evaluation function of a champion-level computer player. From intensive experiments, we found that the PUCT algorithm works very well especially with a good evaluation function and that the value function has more importance than the policy function in the PUCT algorithm.
{"title":"Empirical Analysis of PUCT Algorithm with Evaluation Functions of Different Quality","authors":"Kiminori Matsuzaki","doi":"10.1109/TAAI.2018.00043","DOIUrl":"https://doi.org/10.1109/TAAI.2018.00043","url":null,"abstract":"Monte-Carlo tree search (MCTS) algorithms play an important role in developing computer players for many games. The performance of MCTS players is often leveraged in combination with offline knowledge, i.e., evaluation functions. In particular, recently AlphaGo and AlphaGo Zero achieved a big success in developing strong computer Go player by combining evaluation functions consisting of deep neural networks with a variant of PUCT (Predictor + UCB applied to trees). The effect of evaluation functions on the strength of MCTS algorithms, however, has not been investigated well, especially in terms of the quality of evaluation functions. In this study, we address this issue and empirically analyze the AlphaGo's PUCT algorithm by using Othello (Reversi) as the target game. We investigate the strength of PUCT players using variants of an existing evaluation function of a champion-level computer player. From intensive experiments, we found that the PUCT algorithm works very well especially with a good evaluation function and that the value function has more importance than the policy function in the PUCT algorithm.","PeriodicalId":211734,"journal":{"name":"2018 Conference on Technologies and Applications of Artificial Intelligence (TAAI)","volume":"25 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123502514","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}
Energy disaggregation is a technique of estimation electricity consumption of individual appliance from an aggre-gated meter. In this paper, we study ladder network [6] and multitask learning on energy disaggregation using auto-encoder architecture. This auto-encoder architecture was proposed fromKelly and Knottenbelt in their recent research work [1]. We used this auto-encoder architecture to the high-ownership appliances, air conditioner, bottle warmer, fridge, television and washing machine, in Taiwan and evaluated the effectiveness of the ladder network and multitask learning via these five appliances. The experimental data set has collected by Institute For InformationIndustry. We expect that this project can promote the industrial development of big data-driven smart energy management inTaiwan.
{"title":"An Empirical Study of Ladder Network and Multitask Learning on Energy Disaggregation in Taiwan","authors":"Fang-Yi Chang, Chun-An Chen, Shou-De Lin","doi":"10.1109/TAAI.2018.00028","DOIUrl":"https://doi.org/10.1109/TAAI.2018.00028","url":null,"abstract":"Energy disaggregation is a technique of estimation electricity consumption of individual appliance from an aggre-gated meter. In this paper, we study ladder network [6] and multitask learning on energy disaggregation using auto-encoder architecture. This auto-encoder architecture was proposed fromKelly and Knottenbelt in their recent research work [1]. We used this auto-encoder architecture to the high-ownership appliances, air conditioner, bottle warmer, fridge, television and washing machine, in Taiwan and evaluated the effectiveness of the ladder network and multitask learning via these five appliances. The experimental data set has collected by Institute For InformationIndustry. We expect that this project can promote the industrial development of big data-driven smart energy management inTaiwan.","PeriodicalId":211734,"journal":{"name":"2018 Conference on Technologies and Applications of Artificial Intelligence (TAAI)","volume":"51 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133830762","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}
Metro systems play an important role in reducing traffic congestion in large cities. In this paper, inspired by the potential impact of weather on passenger flow, we have developed an RNN-based model for metro passenger flow prediction with historical passenger flow data, the corresponding temporal data and weather data. A case study of passenger flow prediction model at Taipei Main Station is performed. The experimental results verify that adding the weather data to construct a passenger flow prediction model is contributory to improve the results.
{"title":"Passenger Flow Prediction Using Weather Data for Metro Systems","authors":"Lijuan Liu, R. Chen, Shunzhi Zhu","doi":"10.1109/TAAI.2018.00024","DOIUrl":"https://doi.org/10.1109/TAAI.2018.00024","url":null,"abstract":"Metro systems play an important role in reducing traffic congestion in large cities. In this paper, inspired by the potential impact of weather on passenger flow, we have developed an RNN-based model for metro passenger flow prediction with historical passenger flow data, the corresponding temporal data and weather data. A case study of passenger flow prediction model at Taipei Main Station is performed. The experimental results verify that adding the weather data to construct a passenger flow prediction model is contributory to improve the results.","PeriodicalId":211734,"journal":{"name":"2018 Conference on Technologies and Applications of Artificial Intelligence (TAAI)","volume":"69 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132051293","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}
Chen-Kai Wang, Hong-Jie Dai, Feng-Duo Wang, E. C. Su
Nowadays, social media is often being used by users to create public messages related to their health. With the increasing number of social media usage, a trend has been observed of users creating posts related to adverse drug reactions (ADR). Mining social media data for these information can be used for pharmacological post-marketing surveillance and monitoring. However, the development of automatic ADR detection systems remains challenging because the corpora compiled from real world social media were usually highly imbalanced resulting in barriers to develop classifiers with reliable performance. In this work, we implemented a variety of imbalanced techniques and compared their performance on two large imbalanced data sets released for the purpose of detecting ADR posts. Comparing with state-of-the-art approaches developed for the two dataset, based on much less features, the developed classifiers with implemented imbalanced classification techniques achieved comparable or even better F-scores.
{"title":"Adverse Drug Reaction Post Classification with Imbalanced Classification Techniques","authors":"Chen-Kai Wang, Hong-Jie Dai, Feng-Duo Wang, E. C. Su","doi":"10.1109/TAAI.2018.00011","DOIUrl":"https://doi.org/10.1109/TAAI.2018.00011","url":null,"abstract":"Nowadays, social media is often being used by users to create public messages related to their health. With the increasing number of social media usage, a trend has been observed of users creating posts related to adverse drug reactions (ADR). Mining social media data for these information can be used for pharmacological post-marketing surveillance and monitoring. However, the development of automatic ADR detection systems remains challenging because the corpora compiled from real world social media were usually highly imbalanced resulting in barriers to develop classifiers with reliable performance. In this work, we implemented a variety of imbalanced techniques and compared their performance on two large imbalanced data sets released for the purpose of detecting ADR posts. Comparing with state-of-the-art approaches developed for the two dataset, based on much less features, the developed classifiers with implemented imbalanced classification techniques achieved comparable or even better F-scores.","PeriodicalId":211734,"journal":{"name":"2018 Conference on Technologies and Applications of Artificial Intelligence (TAAI)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121063843","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}
Chun-Wei Lin, Yuyu Zhang, Chun-Hao Chen, J. Wu, Chien-Ming Chen, T. Hong
In this paper, a multi-objective particle swarm optimization (MOPSO)-based framework is presented to find the multiple solutions rather than a single one. The presented grid-based algorithm is used to assign the probability of the non-dominated solution for next iteration. Based on the designed algorithm, it is unnecessary to pre-define the weights of the side effects for evaluation but the non-dominated solutions can be discovered as an alternative way for data sanitization. Extensive experiments are carried on two datasets to show that the designed grid-based algorithm achieves good performance than the traditional single-objective evolution algorithms.
{"title":"A Multiple Objective PSO-Based Approach for Data Sanitization","authors":"Chun-Wei Lin, Yuyu Zhang, Chun-Hao Chen, J. Wu, Chien-Ming Chen, T. Hong","doi":"10.1109/TAAI.2018.00039","DOIUrl":"https://doi.org/10.1109/TAAI.2018.00039","url":null,"abstract":"In this paper, a multi-objective particle swarm optimization (MOPSO)-based framework is presented to find the multiple solutions rather than a single one. The presented grid-based algorithm is used to assign the probability of the non-dominated solution for next iteration. Based on the designed algorithm, it is unnecessary to pre-define the weights of the side effects for evaluation but the non-dominated solutions can be discovered as an alternative way for data sanitization. Extensive experiments are carried on two datasets to show that the designed grid-based algorithm achieves good performance than the traditional single-objective evolution algorithms.","PeriodicalId":211734,"journal":{"name":"2018 Conference on Technologies and Applications of Artificial Intelligence (TAAI)","volume":"86 10 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130995702","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}
Data clustering is a well-known data mining approach that usually used to minimizes the intra distance but maximizes inter distance of each data center. The cluster problem has been proved to be an NP-hard problem. In this paper, a hybrid algorithm based on Whole optimization algorithm (WOA) and Crow search algorithm (CSA) is proposed, namely HWCA. The HWCA algorithm has the advantages of the search strategy of the WOA and CSA. In addition to, there are two operators used to improve the quality of solution, namely hybrid individual operator and enhance diversity operator. The hybrid individual operator is used to exchanges individuals from the WOA and CSA systems by using the roulette wheel approach. In other hand, the HWCA performs enhance diversity operator to improve the quality of each system. More over, the HWCA is incorporated with center optimization strategy to enhance diversity of each system. In the performance evaluation, the proposed MPGO algorithm was comparison WOA and CSA algorithm with six well-known UCI benchmarks. The results show that the proposed algorithm has a higher measure of accuracy rate with comparison algorithms.
{"title":"A Whole Crow Search Algorithm for Solving Data Clustering","authors":"Ze-Xue Wu, Ko-Wei Huang, A. S. Girsang","doi":"10.1109/TAAI.2018.00040","DOIUrl":"https://doi.org/10.1109/TAAI.2018.00040","url":null,"abstract":"Data clustering is a well-known data mining approach that usually used to minimizes the intra distance but maximizes inter distance of each data center. The cluster problem has been proved to be an NP-hard problem. In this paper, a hybrid algorithm based on Whole optimization algorithm (WOA) and Crow search algorithm (CSA) is proposed, namely HWCA. The HWCA algorithm has the advantages of the search strategy of the WOA and CSA. In addition to, there are two operators used to improve the quality of solution, namely hybrid individual operator and enhance diversity operator. The hybrid individual operator is used to exchanges individuals from the WOA and CSA systems by using the roulette wheel approach. In other hand, the HWCA performs enhance diversity operator to improve the quality of each system. More over, the HWCA is incorporated with center optimization strategy to enhance diversity of each system. In the performance evaluation, the proposed MPGO algorithm was comparison WOA and CSA algorithm with six well-known UCI benchmarks. The results show that the proposed algorithm has a higher measure of accuracy rate with comparison algorithms.","PeriodicalId":211734,"journal":{"name":"2018 Conference on Technologies and Applications of Artificial Intelligence (TAAI)","volume":"26 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127210172","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}
Reinforcement Learning is a kind of machine learning method which learns through agents' interaction with the environment. Deep Q-Network (DQN), which is a model of reinforcement learning based on deep neural networks, succeeded in learning human-level control policies on various kinds of Atari 2600 games with image pixel inputs. Because an input of DQN is the game frames of the last four steps, DQN had difficulty on mastering such games that need to remember events earlier than four steps in the past. To alleviate the problem, Deep Recurrent Q-Network (DRQN) and Deep Attention Recurrent Q-Network (DARQN) were proposed. In DRQN, the first fully-connected layer just after convolutional layers is replaced with an LSTM to incorporate past information. DARQN is a model with visual attention mechanisms on top of DRQN. We propose two new reinforcement learning models: Deep Recurrent Q-Network with Truncated History (T-DRQN) and Deep Attention Recurrent Q-Network with Truncated History (T-DARQN). T-DRQN uses a truncated history so that we can control the length of history to be considered. T-DARQN is a model with visual attention mechanism on top of T-DRQN. Experiments of our models on six games of Atari 2600 shows a level of performance between DQN and D(A) RQN. Furthermore, results show the necessity of using past information with a truncated length, rather than using only the current information or all of the past information.
{"title":"Deep Recurrent Q-Network with Truncated History","authors":"Hyunwoo Oh, Tomoyuki Kaneko","doi":"10.1109/TAAI.2018.00017","DOIUrl":"https://doi.org/10.1109/TAAI.2018.00017","url":null,"abstract":"Reinforcement Learning is a kind of machine learning method which learns through agents' interaction with the environment. Deep Q-Network (DQN), which is a model of reinforcement learning based on deep neural networks, succeeded in learning human-level control policies on various kinds of Atari 2600 games with image pixel inputs. Because an input of DQN is the game frames of the last four steps, DQN had difficulty on mastering such games that need to remember events earlier than four steps in the past. To alleviate the problem, Deep Recurrent Q-Network (DRQN) and Deep Attention Recurrent Q-Network (DARQN) were proposed. In DRQN, the first fully-connected layer just after convolutional layers is replaced with an LSTM to incorporate past information. DARQN is a model with visual attention mechanisms on top of DRQN. We propose two new reinforcement learning models: Deep Recurrent Q-Network with Truncated History (T-DRQN) and Deep Attention Recurrent Q-Network with Truncated History (T-DARQN). T-DRQN uses a truncated history so that we can control the length of history to be considered. T-DARQN is a model with visual attention mechanism on top of T-DRQN. Experiments of our models on six games of Atari 2600 shows a level of performance between DQN and D(A) RQN. Furthermore, results show the necessity of using past information with a truncated length, rather than using only the current information or all of the past information.","PeriodicalId":211734,"journal":{"name":"2018 Conference on Technologies and Applications of Artificial Intelligence (TAAI)","volume":"39 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116088324","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：