Pub Date : 2023-02-20DOI: 10.1109/ICAIIC57133.2023.10067005
Heesung Yang, Hyeyoung Park
Near-duplicate image detection is a task to find clusters of images that are considered to be the same pictures in human view. This is important in image recommendation systems, because when the systems recommend candidate images, redundancies of retrieved candidate images need to be avoided. In addition, in the era of big-data where image data is overflowing, its importance in terms of saving storage resources further increases. In this paper, we propose a robust model for detecting various types of near-duplicate images by integrating four different detection modules, where we use multiple image feature extractors such as Gabor filter and deep networks. The four modules are then integrated to conduct the multivariate log-likelihood ratio test for detecting duplication. Through computational experiments, we confirmed that our method reaches state-of-the-art performance.
{"title":"An Integrated Approach to Near-duplicate Image Detection","authors":"Heesung Yang, Hyeyoung Park","doi":"10.1109/ICAIIC57133.2023.10067005","DOIUrl":"https://doi.org/10.1109/ICAIIC57133.2023.10067005","url":null,"abstract":"Near-duplicate image detection is a task to find clusters of images that are considered to be the same pictures in human view. This is important in image recommendation systems, because when the systems recommend candidate images, redundancies of retrieved candidate images need to be avoided. In addition, in the era of big-data where image data is overflowing, its importance in terms of saving storage resources further increases. In this paper, we propose a robust model for detecting various types of near-duplicate images by integrating four different detection modules, where we use multiple image feature extractors such as Gabor filter and deep networks. The four modules are then integrated to conduct the multivariate log-likelihood ratio test for detecting duplication. Through computational experiments, we confirmed that our method reaches state-of-the-art performance.","PeriodicalId":105769,"journal":{"name":"2023 International Conference on Artificial Intelligence in Information and Communication (ICAIIC)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2023-02-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126108122","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 : 2023-02-20DOI: 10.1109/ICAIIC57133.2023.10067067
Seung-Jae Kim, H. Yoe, Meong-hun Lee
Every year, poultry farms suffer great damage from avian influenza outbreaks. The outbreak of avian influenza lowers the egg production rate and has a great impact on the market price. Countries around the world are working simultaneously to prevent the spread of avian influenza. However, despite these efforts, avian influenza still outbreaks every year, resulting in large-scale deaths of chickens. Therefore, in this paper, we propose a disease detection system based on K-Nearest Neighbor Algorithm to prevent large-scale spread of avian influenza. We used decrease in feed intake and a decrease in egg laying rate, which are the main symptoms of avian influenza when it outbreaks, as the standard data for the system's decision. If avian influenza is suspected according to the data analysis result, a push message is sent to the farmer's cell phone, and the farmer checks the information on the area suspected of avian influenza through the application linked with the system and transmits it to the server of the national livestock quarantine system. This is how the system is designed to work. Through this disease detection system, we expect that it will be possible to prevent the spread of avian influenza to the surrounding areas and neighboring farms in advance and to contribute to preventing damage to farms.
{"title":"Design of poultry farm disease detection system based on K-Nearest Neighbor Algorithm","authors":"Seung-Jae Kim, H. Yoe, Meong-hun Lee","doi":"10.1109/ICAIIC57133.2023.10067067","DOIUrl":"https://doi.org/10.1109/ICAIIC57133.2023.10067067","url":null,"abstract":"Every year, poultry farms suffer great damage from avian influenza outbreaks. The outbreak of avian influenza lowers the egg production rate and has a great impact on the market price. Countries around the world are working simultaneously to prevent the spread of avian influenza. However, despite these efforts, avian influenza still outbreaks every year, resulting in large-scale deaths of chickens. Therefore, in this paper, we propose a disease detection system based on K-Nearest Neighbor Algorithm to prevent large-scale spread of avian influenza. We used decrease in feed intake and a decrease in egg laying rate, which are the main symptoms of avian influenza when it outbreaks, as the standard data for the system's decision. If avian influenza is suspected according to the data analysis result, a push message is sent to the farmer's cell phone, and the farmer checks the information on the area suspected of avian influenza through the application linked with the system and transmits it to the server of the national livestock quarantine system. This is how the system is designed to work. Through this disease detection system, we expect that it will be possible to prevent the spread of avian influenza to the surrounding areas and neighboring farms in advance and to contribute to preventing damage to farms.","PeriodicalId":105769,"journal":{"name":"2023 International Conference on Artificial Intelligence in Information and Communication (ICAIIC)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2023-02-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132421250","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 : 2023-02-20DOI: 10.1109/ICAIIC57133.2023.10067072
Ryosuke Wakamoto, Shingo Mabu
In the medical field, research on computer-aided diagnosis using machine learning has been actively conducted. While machine learning can achieve high accuracy by collecting a large amount of data, low interpretability of machine learning is an important issue for achieving practical use in the medical field, where missing a disease may lead to fatal results. In this paper, we propose an anomaly detection method that takes the interpretability into account for diagnosing lung sounds. Furthermore, the proposed method incorporates the context information included in the sound data in the machine learning-based anomaly detection method to improve the detection performance while maintaining the interpretability of the detection results.
{"title":"Interpretable Anomaly Detection for Lung Sounds Using Topology","authors":"Ryosuke Wakamoto, Shingo Mabu","doi":"10.1109/ICAIIC57133.2023.10067072","DOIUrl":"https://doi.org/10.1109/ICAIIC57133.2023.10067072","url":null,"abstract":"In the medical field, research on computer-aided diagnosis using machine learning has been actively conducted. While machine learning can achieve high accuracy by collecting a large amount of data, low interpretability of machine learning is an important issue for achieving practical use in the medical field, where missing a disease may lead to fatal results. In this paper, we propose an anomaly detection method that takes the interpretability into account for diagnosing lung sounds. Furthermore, the proposed method incorporates the context information included in the sound data in the machine learning-based anomaly detection method to improve the detection performance while maintaining the interpretability of the detection results.","PeriodicalId":105769,"journal":{"name":"2023 International Conference on Artificial Intelligence in Information and Communication (ICAIIC)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2023-02-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"134163022","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 : 2023-02-20DOI: 10.1109/ICAIIC57133.2023.10067112
Hyojeong Seo, Dong Seog Han
For safe driving, it is essential to accept reliable information from recognition sensors. In this paper, we present a deep learning model that classifies whether radar signals coming in are normal or abnormal. The abnormal signal is defined as noise from the radar and all signals received when the radar fails or is in trouble. It is difficult to determine whether reflected signals are normal or not based only on radar data. Therefore, the camera and radar sensors are used together, considering the radar cross section (RCS) distribution varies by the angle and distance of the object. The proposed model uses data received from camera and radar sensors to determine the normality of object signals. The model shows an accuracy of 96.24%. Through the results of this study, the reliability of radar signals can be determined in the actual driving environment, thereby ensuring the safety of vehicles and pedestrians.
{"title":"Radar Signal Abnormal Point Classification based on Camera-Radar Sensor Fusion","authors":"Hyojeong Seo, Dong Seog Han","doi":"10.1109/ICAIIC57133.2023.10067112","DOIUrl":"https://doi.org/10.1109/ICAIIC57133.2023.10067112","url":null,"abstract":"For safe driving, it is essential to accept reliable information from recognition sensors. In this paper, we present a deep learning model that classifies whether radar signals coming in are normal or abnormal. The abnormal signal is defined as noise from the radar and all signals received when the radar fails or is in trouble. It is difficult to determine whether reflected signals are normal or not based only on radar data. Therefore, the camera and radar sensors are used together, considering the radar cross section (RCS) distribution varies by the angle and distance of the object. The proposed model uses data received from camera and radar sensors to determine the normality of object signals. The model shows an accuracy of 96.24%. Through the results of this study, the reliability of radar signals can be determined in the actual driving environment, thereby ensuring the safety of vehicles and pedestrians.","PeriodicalId":105769,"journal":{"name":"2023 International Conference on Artificial Intelligence in Information and Communication (ICAIIC)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2023-02-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"134397515","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 : 2023-02-20DOI: 10.1109/ICAIIC57133.2023.10067009
Waqas Khalid, M. A. U. Rehman, Trinh Van Chien, Heejung Yu
Simultaneously transmitting and reflecting reconfig-urable intelligent surfaces (STAR-RISs) provide both transmitting and reflecting signals. The combination of STAR-RIS and non-orthogonal multiple access (NOMA) provides higher performance gains. In this paper, we evaluate NOMA downlink transmission with STAR-RIS under phase error and transceiver hardware impairment. We exploit the statistical properties of the effective channel power and evaluate the ergodic rate behaviors for ideal and non-ideal STAR-RIS-NOMA systems. The numerical results confirm the accuracy of the analytical analysis and demonstrate the selection of the system parameters.
{"title":"Simultaneously Transmitting and Reflecting-Reconfigurable Intelligent Surfaces with Hardware Impairment and Phase Error","authors":"Waqas Khalid, M. A. U. Rehman, Trinh Van Chien, Heejung Yu","doi":"10.1109/ICAIIC57133.2023.10067009","DOIUrl":"https://doi.org/10.1109/ICAIIC57133.2023.10067009","url":null,"abstract":"Simultaneously transmitting and reflecting reconfig-urable intelligent surfaces (STAR-RISs) provide both transmitting and reflecting signals. The combination of STAR-RIS and non-orthogonal multiple access (NOMA) provides higher performance gains. In this paper, we evaluate NOMA downlink transmission with STAR-RIS under phase error and transceiver hardware impairment. We exploit the statistical properties of the effective channel power and evaluate the ergodic rate behaviors for ideal and non-ideal STAR-RIS-NOMA systems. The numerical results confirm the accuracy of the analytical analysis and demonstrate the selection of the system parameters.","PeriodicalId":105769,"journal":{"name":"2023 International Conference on Artificial Intelligence in Information and Communication (ICAIIC)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2023-02-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130846854","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 : 2023-02-20DOI: 10.1109/ICAIIC57133.2023.10067068
Neeraj Kumar, A. Narang
In this paper, we propose Generalized Spatio-Temporal Adaptive Normalization (GSTAN) Framework for Generative Adversarial and Deep Learning Inference Architectures. By leveraging higher-order derivatives based temporal feature maps along with spatial feature map, our normalization approach leads to: (a) efficient generation of high-quality videos with better details and enhanced temporal coherence, and, (b) higher accuracy inference on multiple tasks. In order to evaluate model generalization, we performed experimental evaluation on multiple tasks including: video to video generation, video segmentation and activity recognition (classify the activity out of 101 activity classes, for a given input video). Detailed experimental analysis over a variety of datasets including CityScape, UCF101 and CK+ demonstrates superior performance of GSTAN and also provides the impact of its various configurations, including parallel GSTAN and sequential GSTAN.
{"title":"Generalized Spatio-Temporal Adaptive Normalization Framework","authors":"Neeraj Kumar, A. Narang","doi":"10.1109/ICAIIC57133.2023.10067068","DOIUrl":"https://doi.org/10.1109/ICAIIC57133.2023.10067068","url":null,"abstract":"In this paper, we propose Generalized Spatio-Temporal Adaptive Normalization (GSTAN) Framework for Generative Adversarial and Deep Learning Inference Architectures. By leveraging higher-order derivatives based temporal feature maps along with spatial feature map, our normalization approach leads to: (a) efficient generation of high-quality videos with better details and enhanced temporal coherence, and, (b) higher accuracy inference on multiple tasks. In order to evaluate model generalization, we performed experimental evaluation on multiple tasks including: video to video generation, video segmentation and activity recognition (classify the activity out of 101 activity classes, for a given input video). Detailed experimental analysis over a variety of datasets including CityScape, UCF101 and CK+ demonstrates superior performance of GSTAN and also provides the impact of its various configurations, including parallel GSTAN and sequential GSTAN.","PeriodicalId":105769,"journal":{"name":"2023 International Conference on Artificial Intelligence in Information and Communication (ICAIIC)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2023-02-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114690552","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 : 2023-02-20DOI: 10.1109/ICAIIC57133.2023.10067091
Sol Lim, Rahma Gantassi, Yonghoon Choi
In a cost-effective peak shaving strategy, clustering and machine learning algorithm can be used to set optimal peak shaving time zone for each load. Energy Storage System (ESS) charge amount is determined with load prediction data through machine learning model, and the peak shaving time zone is adjusted flexibly according to load patterns for each cluster. It is possible to prevent ESS from being overcharged or undercharged through load prediction. In addition, rather than applying peak shaving collectively at the on-peak time, efficient operation of the power grid can be expected by adjusting the time zone flexibly for each power usage pattern. The effectiveness of the proposed system model is to be proved through changes in electricity cost depending on whether it is introduced or not.
{"title":"Cost-Effective Peak Shaving Strategy Based on Clustering and XGBoost Algorithm","authors":"Sol Lim, Rahma Gantassi, Yonghoon Choi","doi":"10.1109/ICAIIC57133.2023.10067091","DOIUrl":"https://doi.org/10.1109/ICAIIC57133.2023.10067091","url":null,"abstract":"In a cost-effective peak shaving strategy, clustering and machine learning algorithm can be used to set optimal peak shaving time zone for each load. Energy Storage System (ESS) charge amount is determined with load prediction data through machine learning model, and the peak shaving time zone is adjusted flexibly according to load patterns for each cluster. It is possible to prevent ESS from being overcharged or undercharged through load prediction. In addition, rather than applying peak shaving collectively at the on-peak time, efficient operation of the power grid can be expected by adjusting the time zone flexibly for each power usage pattern. The effectiveness of the proposed system model is to be proved through changes in electricity cost depending on whether it is introduced or not.","PeriodicalId":105769,"journal":{"name":"2023 International Conference on Artificial Intelligence in Information and Communication (ICAIIC)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2023-02-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116490732","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 : 2023-02-20DOI: 10.1109/ICAIIC57133.2023.10067039
Wrucha Nanal, M. Hajiarbabi
Image Captioning aspires to achieve a description of images with machines as a combination of Computer Vision (CV) and Natural Language Processing (NLP) fields. The current state of the art for image captioning use the Attention-based Encoder-Decoder model. The Attention-based model uses an ‘Attention mechanism’ that focuses on a particular section of the image to generate its corresponding caption word. The NLP side of this model uses Long Short-Term Memory (LSTM) for word generation. Attention-based models did not emphasize the relative arrangement of words in a caption thereby, ignoring the context of the sentence. Inspired by the versatility of Transformers in NLP, this work tries to utilise its architecture features for the Image Captioning use case. This work also makes use of a pretrained Bidirectional Encoder Representation of Transformer (BERT) which generates a contextually rich embedding of a caption. The Multi-Head Attention of the Transformer establishes a strong correlation between the image and contextually aware caption. This experiment is performed on the Remote Sensing Image Captioning Dataset. The results of the model are evaluated using NLP evaluation metrics such as Bilingual Evaluation Understudy 1–4 (BLEU), Metric for Evaluation of Translation with Explicit ORdering (METEOR) and Recall-Oriented Understudy for Gisting Evaluation (ROUGE). The proposed model shows better results for a few of the metrics.
{"title":"Captioning Remote Sensing Images Using Transformer Architecture","authors":"Wrucha Nanal, M. Hajiarbabi","doi":"10.1109/ICAIIC57133.2023.10067039","DOIUrl":"https://doi.org/10.1109/ICAIIC57133.2023.10067039","url":null,"abstract":"Image Captioning aspires to achieve a description of images with machines as a combination of Computer Vision (CV) and Natural Language Processing (NLP) fields. The current state of the art for image captioning use the Attention-based Encoder-Decoder model. The Attention-based model uses an ‘Attention mechanism’ that focuses on a particular section of the image to generate its corresponding caption word. The NLP side of this model uses Long Short-Term Memory (LSTM) for word generation. Attention-based models did not emphasize the relative arrangement of words in a caption thereby, ignoring the context of the sentence. Inspired by the versatility of Transformers in NLP, this work tries to utilise its architecture features for the Image Captioning use case. This work also makes use of a pretrained Bidirectional Encoder Representation of Transformer (BERT) which generates a contextually rich embedding of a caption. The Multi-Head Attention of the Transformer establishes a strong correlation between the image and contextually aware caption. This experiment is performed on the Remote Sensing Image Captioning Dataset. The results of the model are evaluated using NLP evaluation metrics such as Bilingual Evaluation Understudy 1–4 (BLEU), Metric for Evaluation of Translation with Explicit ORdering (METEOR) and Recall-Oriented Understudy for Gisting Evaluation (ROUGE). The proposed model shows better results for a few of the metrics.","PeriodicalId":105769,"journal":{"name":"2023 International Conference on Artificial Intelligence in Information and Communication (ICAIIC)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2023-02-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123530105","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 : 2023-02-20DOI: 10.1109/ICAIIC57133.2023.10066968
Joon-Il Cho, Soorim Yang, Jae-Hoon Kim
Counting the number of packaged objects is a process that the fulfillment center must go through before shipping it to the customer. With the development of computer vision and deep learning, many studies are being conducted to recognize objects and determine quantity using cameras, but it is impossible to identify objects in a sealed package due to the occlusion problem of not seeing objects that are occluded. Among the nondestructive inspection methods using wavelengths that use penetrating properties, radar that is harmless to the human body does not require contact with objects, and uses low-power radio frequency is the most suitable for counting objects. In this paper, we propose a system that achieves 99.33% counting accuracy of packed objects by removing background noise of radio frequency measured by 60GHz pulsed coherent radar.
{"title":"PCR Radar-Based Counting System for Packaged Objects","authors":"Joon-Il Cho, Soorim Yang, Jae-Hoon Kim","doi":"10.1109/ICAIIC57133.2023.10066968","DOIUrl":"https://doi.org/10.1109/ICAIIC57133.2023.10066968","url":null,"abstract":"Counting the number of packaged objects is a process that the fulfillment center must go through before shipping it to the customer. With the development of computer vision and deep learning, many studies are being conducted to recognize objects and determine quantity using cameras, but it is impossible to identify objects in a sealed package due to the occlusion problem of not seeing objects that are occluded. Among the nondestructive inspection methods using wavelengths that use penetrating properties, radar that is harmless to the human body does not require contact with objects, and uses low-power radio frequency is the most suitable for counting objects. In this paper, we propose a system that achieves 99.33% counting accuracy of packed objects by removing background noise of radio frequency measured by 60GHz pulsed coherent radar.","PeriodicalId":105769,"journal":{"name":"2023 International Conference on Artificial Intelligence in Information and Communication (ICAIIC)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2023-02-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122129502","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 : 2023-02-20DOI: 10.1109/ICAIIC57133.2023.10067098
Taeho Lee, Junhee Seok
Multi-task learning (MTL) is a problem that must be applied in modern recommendation systems and is just as difficult. In the recent e-commerce advertising market, it is necessary to be able to predict not only the probability of users clicking, but also the probability of conversion and purchase. By predicting multi-task, it is possible to increase the accuracy of each task and optimize advertisements for various goals of advertisers. Traditional conversion rate (CVR) prediction models have difficulty learning because the number of conversions is too small compared to the total number of impressions. This problem is called a data sparsity (DS) problem. Another problem is that CVR models trained with samples of clicked impressions infer on samples of all impressions. This problem is called a sample selection bias (SSB) problem. This paper is a summary of the various solutions and current limitations and further directions about solving sample selection bias problem and data sparsity problem.
{"title":"Multi Task Learning: A Survey and Future Directions","authors":"Taeho Lee, Junhee Seok","doi":"10.1109/ICAIIC57133.2023.10067098","DOIUrl":"https://doi.org/10.1109/ICAIIC57133.2023.10067098","url":null,"abstract":"Multi-task learning (MTL) is a problem that must be applied in modern recommendation systems and is just as difficult. In the recent e-commerce advertising market, it is necessary to be able to predict not only the probability of users clicking, but also the probability of conversion and purchase. By predicting multi-task, it is possible to increase the accuracy of each task and optimize advertisements for various goals of advertisers. Traditional conversion rate (CVR) prediction models have difficulty learning because the number of conversions is too small compared to the total number of impressions. This problem is called a data sparsity (DS) problem. Another problem is that CVR models trained with samples of clicked impressions infer on samples of all impressions. This problem is called a sample selection bias (SSB) problem. This paper is a summary of the various solutions and current limitations and further directions about solving sample selection bias problem and data sparsity problem.","PeriodicalId":105769,"journal":{"name":"2023 International Conference on Artificial Intelligence in Information and Communication (ICAIIC)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2023-02-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124024429","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：