Pub Date : 2023-09-29DOI: 10.1007/s40747-023-01244-8
Zheng Chen, Zhejun Liu
Abstract In this study, we primarily aim to address the exposure bias issue in long text generation intrinsic to statistical language models. We propose a sentence-level heuristic tree search algorithm, specially tailored for long text generation, to mitigate the problem by managing generated texts in a tree structure and curbing the compounding of biases. Our algorithm utilizes two pre-trained language models, an auto-regressive model for generating new sentences and an auto-encoder model for evaluating sentence quality. These models work in tandem to perform four critical operations: expanding the text tree with new sentences, evaluating the quality of the additions, sampling potential unfinished text fragments for further generation, and pruning leaf nodes deemed unpromising. This iterative process continues until a pre-defined number of [EOS] tokens are produced, at which point we select the highest-scoring completed text as our final output. Moreover, we pioneer two novel token-level decoding techniques—nucleus sampling with temperature and diverse beam search with sampling. These methods, integrated with our sentence-level search algorithm, aim to improve the consistency and diversity of text generation. Experimental results, both automated measures (including Jaccard similarity, Word2vec similarity, and unique word ratio) and human evaluations (assessing consistency, fluency, and rhetorical skills), conclusively demonstrate that our approach considerably enhances the quality of machine-generated long-form text. Through this research, we aim to inspire further innovations in sentence-level search-based text generation algorithms.
{"title":"Sentence-level heuristic tree search for long text generation","authors":"Zheng Chen, Zhejun Liu","doi":"10.1007/s40747-023-01244-8","DOIUrl":"https://doi.org/10.1007/s40747-023-01244-8","url":null,"abstract":"Abstract In this study, we primarily aim to address the exposure bias issue in long text generation intrinsic to statistical language models. We propose a sentence-level heuristic tree search algorithm, specially tailored for long text generation, to mitigate the problem by managing generated texts in a tree structure and curbing the compounding of biases. Our algorithm utilizes two pre-trained language models, an auto-regressive model for generating new sentences and an auto-encoder model for evaluating sentence quality. These models work in tandem to perform four critical operations: expanding the text tree with new sentences, evaluating the quality of the additions, sampling potential unfinished text fragments for further generation, and pruning leaf nodes deemed unpromising. This iterative process continues until a pre-defined number of [EOS] tokens are produced, at which point we select the highest-scoring completed text as our final output. Moreover, we pioneer two novel token-level decoding techniques—nucleus sampling with temperature and diverse beam search with sampling. These methods, integrated with our sentence-level search algorithm, aim to improve the consistency and diversity of text generation. Experimental results, both automated measures (including Jaccard similarity, Word2vec similarity, and unique word ratio) and human evaluations (assessing consistency, fluency, and rhetorical skills), conclusively demonstrate that our approach considerably enhances the quality of machine-generated long-form text. Through this research, we aim to inspire further innovations in sentence-level search-based text generation algorithms.","PeriodicalId":10524,"journal":{"name":"Complex & Intelligent Systems","volume":"23 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-09-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"135200387","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 : 2023-09-28DOI: 10.1007/s40747-023-01228-8
Yue Gao, Xiangling Fu, Xien Liu, Ji Wu
Abstract Graph-based neural networks and unsupervised pre-trained models are both cutting-edge text representation methods, given their outstanding ability to capture global information and contextualized information, respectively. However, both representation methods meet obstacles to further performance improvements. On one hand, graph-based neural networks lack knowledge orientation to guide textual interpretation during global information interaction. On the other hand, unsupervised pre-trained models imply rich semantic and syntactic knowledge which lacks sufficient induction and expression. Therefore, how to effectively integrate graph-based global information and unsupervised contextualized semantic and syntactic information to achieve better text representation is an important issue pending for solution. In this paper, we propose a representation method that deeply integrates Unsupervised Semantics and Syntax into heterogeneous Graphs (USS-Graph) for inductive text classification. By constructing a heterogeneous graph whose edges and nodes are totally generated by knowledge from unsupervised pre-trained models, USS-Graph can harmonize the two perspectives of information under a bidirectionally weighted graph structure and thereby realizing the intra-fusion of graph-based global information and unsupervised contextualized semantic and syntactic information. Based on USS-Graph, we also propose a series of optimization measures to further improve the knowledge integration and representation performance. Extensive experiments conducted on benchmark datasets show that USS-Graph consistently achieves state-of-the-art performances on inductive text classification tasks. Additionally, extended experiments are conducted to deeply analyze the characteristics of USS-Graph and the effectiveness of our proposed optimization measures for further knowledge integration and information complementation.
{"title":"Deeply integrating unsupervised semantics and syntax into heterogeneous graphs for inductive text classification","authors":"Yue Gao, Xiangling Fu, Xien Liu, Ji Wu","doi":"10.1007/s40747-023-01228-8","DOIUrl":"https://doi.org/10.1007/s40747-023-01228-8","url":null,"abstract":"Abstract Graph-based neural networks and unsupervised pre-trained models are both cutting-edge text representation methods, given their outstanding ability to capture global information and contextualized information, respectively. However, both representation methods meet obstacles to further performance improvements. On one hand, graph-based neural networks lack knowledge orientation to guide textual interpretation during global information interaction. On the other hand, unsupervised pre-trained models imply rich semantic and syntactic knowledge which lacks sufficient induction and expression. Therefore, how to effectively integrate graph-based global information and unsupervised contextualized semantic and syntactic information to achieve better text representation is an important issue pending for solution. In this paper, we propose a representation method that deeply integrates Unsupervised Semantics and Syntax into heterogeneous Graphs (USS-Graph) for inductive text classification. By constructing a heterogeneous graph whose edges and nodes are totally generated by knowledge from unsupervised pre-trained models, USS-Graph can harmonize the two perspectives of information under a bidirectionally weighted graph structure and thereby realizing the intra-fusion of graph-based global information and unsupervised contextualized semantic and syntactic information. Based on USS-Graph, we also propose a series of optimization measures to further improve the knowledge integration and representation performance. Extensive experiments conducted on benchmark datasets show that USS-Graph consistently achieves state-of-the-art performances on inductive text classification tasks. Additionally, extended experiments are conducted to deeply analyze the characteristics of USS-Graph and the effectiveness of our proposed optimization measures for further knowledge integration and information complementation.","PeriodicalId":10524,"journal":{"name":"Complex & Intelligent Systems","volume":"31 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-09-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"135343354","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 : 2023-09-28DOI: 10.1007/s40747-023-01226-w
Tianwei Yan, Xiang Zhang, Zhigang Luo
Abstract Distantly supervised relation extraction is an automatically annotating method for large corpora by classifying a bound of sentences with two same entities and the relation. Recent works exploit sound performance by adopting contrastive learning to efficiently obtain instance representations under the multi-instance learning framework. Though these methods weaken the impact of noisy labels, it ignores the long-tail distribution problem in distantly supervised sets and fails to capture the mutual information of different parts. We are thus motivated to tackle these issues and establishing a long-tail awareness contrastive learning method for efficiently utilizing the long-tail data. Our model treats major and tail parts differently by adopting hyper-augmentation strategies. Moreover, the model provides various views by constructing novel positive and negative pairs in contrastive learning for gaining a better representation between different parts. The experimental results on the NYT10 dataset demonstrate our model surpasses the existing SOTA by more than 2.61% AUC score on relation extraction. In manual evaluation datasets including NYT10m and Wiki20m, our method obtains competitive results by achieving 59.42% and 79.19% AUC scores on relation extraction, respectively. Extensive discussions further confirm the effectiveness of our approach.
{"title":"LTACL: long-tail awareness contrastive learning for distantly supervised relation extraction","authors":"Tianwei Yan, Xiang Zhang, Zhigang Luo","doi":"10.1007/s40747-023-01226-w","DOIUrl":"https://doi.org/10.1007/s40747-023-01226-w","url":null,"abstract":"Abstract Distantly supervised relation extraction is an automatically annotating method for large corpora by classifying a bound of sentences with two same entities and the relation. Recent works exploit sound performance by adopting contrastive learning to efficiently obtain instance representations under the multi-instance learning framework. Though these methods weaken the impact of noisy labels, it ignores the long-tail distribution problem in distantly supervised sets and fails to capture the mutual information of different parts. We are thus motivated to tackle these issues and establishing a long-tail awareness contrastive learning method for efficiently utilizing the long-tail data. Our model treats major and tail parts differently by adopting hyper-augmentation strategies. Moreover, the model provides various views by constructing novel positive and negative pairs in contrastive learning for gaining a better representation between different parts. The experimental results on the NYT10 dataset demonstrate our model surpasses the existing SOTA by more than 2.61% AUC score on relation extraction. In manual evaluation datasets including NYT10m and Wiki20m, our method obtains competitive results by achieving 59.42% and 79.19% AUC scores on relation extraction, respectively. Extensive discussions further confirm the effectiveness of our approach.","PeriodicalId":10524,"journal":{"name":"Complex & Intelligent Systems","volume":"23 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-09-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"135344553","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 : 2023-09-27DOI: 10.1007/s40747-023-01231-z
Kuangrong Hao, Chenwei Zhao, Xiaoyan Liu
Abstract Communication recovery is necessary for rescue and reconstruction scenarios including earthquakes, typhoons, floods, etc. The rapid and stable communication link can provide efficient victims’ real-time information for the rescue process. However, traditional centralized communication links cannot traverse the further victims with information-sharing requirements. And the even communication link distribution leads to a load burden on the crowded victim area. Thus, we propose a three-layer architecture consisting of the emergency communication vehicle, backbone links, and branch links to rapidly recover communication via mobile robots. Then, considering victims’ distribution, an improved MaxMin distance algorithm is presented as the basis of robot dispatch. The relay probability of the link is also estimated with closed formulae. Finally, simulation results verify that our proposed algorithm can recover communication with lower delay and higher packet delivery ratio.
{"title":"A robot-assisted adaptive communication recovery method in disaster scenarios","authors":"Kuangrong Hao, Chenwei Zhao, Xiaoyan Liu","doi":"10.1007/s40747-023-01231-z","DOIUrl":"https://doi.org/10.1007/s40747-023-01231-z","url":null,"abstract":"Abstract Communication recovery is necessary for rescue and reconstruction scenarios including earthquakes, typhoons, floods, etc. The rapid and stable communication link can provide efficient victims’ real-time information for the rescue process. However, traditional centralized communication links cannot traverse the further victims with information-sharing requirements. And the even communication link distribution leads to a load burden on the crowded victim area. Thus, we propose a three-layer architecture consisting of the emergency communication vehicle, backbone links, and branch links to rapidly recover communication via mobile robots. Then, considering victims’ distribution, an improved MaxMin distance algorithm is presented as the basis of robot dispatch. The relay probability of the link is also estimated with closed formulae. Finally, simulation results verify that our proposed algorithm can recover communication with lower delay and higher packet delivery ratio.","PeriodicalId":10524,"journal":{"name":"Complex & Intelligent Systems","volume":"22 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-09-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"135477402","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 : 2023-09-26DOI: 10.1007/s40747-023-01236-8
Minming Gu, Kaiyu Chen, Zhixiang Chen
Abstract Dangerous driving behavior is a major contributing factor to road traffic accidents. Identifying and intervening in drivers’ unsafe driving behaviors is thus crucial for preventing accidents and ensuring road safety. However, many of the existing methods for monitoring drivers’ behaviors rely on computer vision technology, which has the potential to invade privacy. This paper proposes a radar-based deep learning method to analyze driver behavior. The method utilizes FMCW radar along with TOF radar to identify five types of driving behavior: normal driving, head up, head twisting, picking up the phone, and dancing to music. The proposed model, called RFDANet, includes two parallel forward propagation channels that are relatively independent of each other. The range-Doppler information from the FMCW radar and the position information from the TOF radar are used as inputs. After feature extraction by CNN, an attention mechanism is introduced into the deep architecture of the branch layer to adjust the weight of different branches. To further recognize driving behavior, LSTM is used. The effectiveness of the proposed method is verified by actual driving data. The results indicate that the average accuracy of each of the five types of driving behavior is 94.5%, which shows the advantage of using the proposed deep learning method. Overall, the experimental results confirm that the proposed method is highly effective for detecting drivers’ behavior.
{"title":"RFDANet: an FMCW and TOF radar fusion approach for driver activity recognition using multi-level attention based CNN and LSTM network","authors":"Minming Gu, Kaiyu Chen, Zhixiang Chen","doi":"10.1007/s40747-023-01236-8","DOIUrl":"https://doi.org/10.1007/s40747-023-01236-8","url":null,"abstract":"Abstract Dangerous driving behavior is a major contributing factor to road traffic accidents. Identifying and intervening in drivers’ unsafe driving behaviors is thus crucial for preventing accidents and ensuring road safety. However, many of the existing methods for monitoring drivers’ behaviors rely on computer vision technology, which has the potential to invade privacy. This paper proposes a radar-based deep learning method to analyze driver behavior. The method utilizes FMCW radar along with TOF radar to identify five types of driving behavior: normal driving, head up, head twisting, picking up the phone, and dancing to music. The proposed model, called RFDANet, includes two parallel forward propagation channels that are relatively independent of each other. The range-Doppler information from the FMCW radar and the position information from the TOF radar are used as inputs. After feature extraction by CNN, an attention mechanism is introduced into the deep architecture of the branch layer to adjust the weight of different branches. To further recognize driving behavior, LSTM is used. The effectiveness of the proposed method is verified by actual driving data. The results indicate that the average accuracy of each of the five types of driving behavior is 94.5%, which shows the advantage of using the proposed deep learning method. Overall, the experimental results confirm that the proposed method is highly effective for detecting drivers’ behavior.","PeriodicalId":10524,"journal":{"name":"Complex & Intelligent Systems","volume":"42 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-09-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"134961081","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 : 2023-09-15DOI: 10.1007/s40747-023-01223-z
Yuanyun Wang, Wenhui Yang, Peng Yin, Jun Wang
Abstract Siamese-based trackers have been widely studied for their high accuracy and speed. Both the feature extraction and feature fusion are two important components in Siamese-based trackers. Siamese-based trackers obtain fine local features by traditional convolution. However, some important channel information and global information are lost when enhancing local features. In the feature fusion process, cross-correlation-based feature fusion between the template and search region feature ignores the global spatial context information and does not make the best of the spatial information. In this paper, to solve the above problem, we design a novel feature extraction sub-network based on batch-free normalization re-parameterization convolution, which scales the features in the channel dimension and increases the receptive field. Richer channel information is obtained and powerful target features are extracted for the feature fusion. Furthermore, we learn a feature fusion network (FFN) based on feature filter. The FFN fuses the template and search region features in a global spatial context to obtain high-quality fused features by enhancing important features and filtering redundant features. By jointly learning the proposed feature extraction sub-network and FFN, the local and global information are fully exploited. Then, we propose a novel tracking algorithm based on the designed feature extraction sub-network and FFN with re-parameterization convolution and feature filter, referred to as RCFT. We evaluate the proposed RCFT tracker and some recent state-of-the-art (SOTA) trackers on OTB100, VOT2018, LaSOT, GOT-10k, UAV123 and the visual-thermal dataset VOT-RGBT2019 datasets, which achieves superior tracking performance with 45 FPS tracking speed.
{"title":"RCFT: re-parameterization convolution and feature filter for object tracking","authors":"Yuanyun Wang, Wenhui Yang, Peng Yin, Jun Wang","doi":"10.1007/s40747-023-01223-z","DOIUrl":"https://doi.org/10.1007/s40747-023-01223-z","url":null,"abstract":"Abstract Siamese-based trackers have been widely studied for their high accuracy and speed. Both the feature extraction and feature fusion are two important components in Siamese-based trackers. Siamese-based trackers obtain fine local features by traditional convolution. However, some important channel information and global information are lost when enhancing local features. In the feature fusion process, cross-correlation-based feature fusion between the template and search region feature ignores the global spatial context information and does not make the best of the spatial information. In this paper, to solve the above problem, we design a novel feature extraction sub-network based on batch-free normalization re-parameterization convolution, which scales the features in the channel dimension and increases the receptive field. Richer channel information is obtained and powerful target features are extracted for the feature fusion. Furthermore, we learn a feature fusion network (FFN) based on feature filter. The FFN fuses the template and search region features in a global spatial context to obtain high-quality fused features by enhancing important features and filtering redundant features. By jointly learning the proposed feature extraction sub-network and FFN, the local and global information are fully exploited. Then, we propose a novel tracking algorithm based on the designed feature extraction sub-network and FFN with re-parameterization convolution and feature filter, referred to as RCFT. We evaluate the proposed RCFT tracker and some recent state-of-the-art (SOTA) trackers on OTB100, VOT2018, LaSOT, GOT-10k, UAV123 and the visual-thermal dataset VOT-RGBT2019 datasets, which achieves superior tracking performance with 45 FPS tracking speed.","PeriodicalId":10524,"journal":{"name":"Complex & Intelligent Systems","volume":"23 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-09-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"135437228","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 : 2023-09-14DOI: 10.1007/s40747-023-01227-9
Zongling Li, Xinjiang Chen, Qizhang Luo, Guohua Wu, Ling Wang
Abstract System disturbances, such as the change of required service durations, the failure of resources, and temporary tasks during the scheduling process of data relay satellite network (DRSN), are difficult to be predicted, which may lead to unsuccessful scheduling of tasks. A high-efficiency and robust DRSN calls for smarter and more flexible disturbances elimination strategies. Here, we unify the above three system disturbances as temporary task arrival and extend the static scheduling model of DRSN. Specifically, we derive and define a scheduling model that unifies the static scheduling and dynamic scheduling processes. Meanwhile, we propose a k -step dynamic scheduling algorithm considering breakpoint transmission ( k -steps-BT) to solve the above model. Based on the principle of backtracking algorithm and search tree, k -steps-BT can eliminate disturbances quickly by rescheduling tasks and can determine the rescheduling scheme when temporary tasks arrive. Finally, extensive experiments are carried out to verify the proposed model and algorithm. The results show that the proposed model and algorithm can significantly improve the task completion rate of dynamic scheduling without drastic adjustments to the static scheduling scheme.
{"title":"Dynamic scheduling method for data relay satellite networks considering hybrid system disturbances","authors":"Zongling Li, Xinjiang Chen, Qizhang Luo, Guohua Wu, Ling Wang","doi":"10.1007/s40747-023-01227-9","DOIUrl":"https://doi.org/10.1007/s40747-023-01227-9","url":null,"abstract":"Abstract System disturbances, such as the change of required service durations, the failure of resources, and temporary tasks during the scheduling process of data relay satellite network (DRSN), are difficult to be predicted, which may lead to unsuccessful scheduling of tasks. A high-efficiency and robust DRSN calls for smarter and more flexible disturbances elimination strategies. Here, we unify the above three system disturbances as temporary task arrival and extend the static scheduling model of DRSN. Specifically, we derive and define a scheduling model that unifies the static scheduling and dynamic scheduling processes. Meanwhile, we propose a k -step dynamic scheduling algorithm considering breakpoint transmission ( k -steps-BT) to solve the above model. Based on the principle of backtracking algorithm and search tree, k -steps-BT can eliminate disturbances quickly by rescheduling tasks and can determine the rescheduling scheme when temporary tasks arrive. Finally, extensive experiments are carried out to verify the proposed model and algorithm. The results show that the proposed model and algorithm can significantly improve the task completion rate of dynamic scheduling without drastic adjustments to the static scheduling scheme.","PeriodicalId":10524,"journal":{"name":"Complex & Intelligent Systems","volume":"8 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-09-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"134910560","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 : 2023-09-13DOI: 10.1007/s40747-023-01218-w
Qi Jia, Jing Guo, Po Yang, Yun Yang
Abstract Human activity recognition (HAR) aims to collect time series through wearable devices to precisely identify specific actions. However, the traditional HAR method ignores the activity variances among individuals, which will cause low generalization when applied to a new individual and indirectly enhance the difficulties of personalized HAR service. In this paper, we fully consider activity divergence among individuals to develop an end-to-end model, the multi-source unsupervised co-transfer network (MUCT), to provide personalized activity recognition for new individuals. We denote the collected data of different individuals as multiple domains and implement deep domain adaptation to align each pair of source and target domains. In addition, we propose a consistent filter that utilizes two heterogeneous classifiers to automatically select high-confidence instances from the target domain to jointly enhance the performance on the target task. The effectiveness and performance of our model are evaluated through comprehensive experiments on two activity recognition benchmarks and a private activity recognition data set (collected by our signal sensors), where our model outperforms traditional transfer learning methods at HAR.
{"title":"A holistic multi-source transfer learning approach using wearable sensors for personalized daily activity recognition","authors":"Qi Jia, Jing Guo, Po Yang, Yun Yang","doi":"10.1007/s40747-023-01218-w","DOIUrl":"https://doi.org/10.1007/s40747-023-01218-w","url":null,"abstract":"Abstract Human activity recognition (HAR) aims to collect time series through wearable devices to precisely identify specific actions. However, the traditional HAR method ignores the activity variances among individuals, which will cause low generalization when applied to a new individual and indirectly enhance the difficulties of personalized HAR service. In this paper, we fully consider activity divergence among individuals to develop an end-to-end model, the multi-source unsupervised co-transfer network (MUCT), to provide personalized activity recognition for new individuals. We denote the collected data of different individuals as multiple domains and implement deep domain adaptation to align each pair of source and target domains. In addition, we propose a consistent filter that utilizes two heterogeneous classifiers to automatically select high-confidence instances from the target domain to jointly enhance the performance on the target task. The effectiveness and performance of our model are evaluated through comprehensive experiments on two activity recognition benchmarks and a private activity recognition data set (collected by our signal sensors), where our model outperforms traditional transfer learning methods at HAR.","PeriodicalId":10524,"journal":{"name":"Complex & Intelligent Systems","volume":"11 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-09-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"135741570","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 : 2023-09-13DOI: 10.1007/s40747-023-01219-9
ShanChen Pang, WenShang Zhao, ShuDong Wang, Lin Zhang, Shuang Wang
Abstract Computer vision has developed rapidly in recent years, invigorating the area of industrial surface defect detection while also providing it with modern perception capabilities. Few-shot learning has emerged as a result of sample size limitations, with MAML framework being the most widely used few-shot learning framework over the past few years that learns concepts from sampled classification tasks, which is considered to have the key advantage of aligning training and testing objectives. Industrial surface defects typically have fewer samples for training, so we propose MAML-based framework: Permute-MAML, which mainly consists of improved MAML framework and neural network model. In this paper, we concentrate on improving MAML framework with respect to its stability and explore a simple procedure: few-shot learning of its evaluation metrics over the whole classification model. The experimental results demonstrate that the proposed framework significantly enhances the stability of MAML framework and achieves comparatively high accuracy in industrial surface defect detection.
{"title":"Permute-MAML: exploring industrial surface defect detection algorithms for few-shot learning","authors":"ShanChen Pang, WenShang Zhao, ShuDong Wang, Lin Zhang, Shuang Wang","doi":"10.1007/s40747-023-01219-9","DOIUrl":"https://doi.org/10.1007/s40747-023-01219-9","url":null,"abstract":"Abstract Computer vision has developed rapidly in recent years, invigorating the area of industrial surface defect detection while also providing it with modern perception capabilities. Few-shot learning has emerged as a result of sample size limitations, with MAML framework being the most widely used few-shot learning framework over the past few years that learns concepts from sampled classification tasks, which is considered to have the key advantage of aligning training and testing objectives. Industrial surface defects typically have fewer samples for training, so we propose MAML-based framework: Permute-MAML, which mainly consists of improved MAML framework and neural network model. In this paper, we concentrate on improving MAML framework with respect to its stability and explore a simple procedure: few-shot learning of its evaluation metrics over the whole classification model. The experimental results demonstrate that the proposed framework significantly enhances the stability of MAML framework and achieves comparatively high accuracy in industrial surface defect detection.","PeriodicalId":10524,"journal":{"name":"Complex & Intelligent Systems","volume":"31 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-09-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"135741085","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}
Abstract Hesitant Fermatean fuzzy sets (HFFS) can characterize the membership degree (MD) and non-membership degree (NMD) of hesitant fuzzy elements in a broader range, which offers superior fuzzy data processing capabilities for addressing complex uncertainty issues. In this research, first, we present the definition of the hesitant Fermatean fuzzy Bonferroni mean operator (HFFBM). Further, with the basic operations of HFFS in Einstein t-norms, the definition and derivation process of the hesitant Fermatean fuzzy Einstein Bonferroni mean operator (HFFEBM) are given. In addition, considering how weights affect decision-making outcomes, the hesitant Fermatean fuzzy weighted Bonferroni mean (HFFWBM) operator and the hesitant Fermatean fuzzy Einstein weighted Bonferroni mean operator (HFFEWBM) are developed. Then, the properties of the operators are discussed. Based on HFFWBM and HFFEWBM operator, a new multi-attribute decision-making (MADM) approach is provided. Finally, we apply the proposed decision-making approach to the case of a depression diagnostic evaluation for three depressed patients. The three patients' diagnosis results confirmed the proposed method's validity and rationality. Through a series of comparative experiments and analyses, the proposed MADM method is an efficient solution for decision-making issues in the hesitant Fermatean fuzzy environment.
{"title":"Hesitant Fermatean fuzzy Bonferroni mean operators for multi-attribute decision-making","authors":"Yibo Wang, Xiuqin Ma, Hongwu Qin, Huanling Sun, Weiyi Wei","doi":"10.1007/s40747-023-01203-3","DOIUrl":"https://doi.org/10.1007/s40747-023-01203-3","url":null,"abstract":"Abstract Hesitant Fermatean fuzzy sets (HFFS) can characterize the membership degree (MD) and non-membership degree (NMD) of hesitant fuzzy elements in a broader range, which offers superior fuzzy data processing capabilities for addressing complex uncertainty issues. In this research, first, we present the definition of the hesitant Fermatean fuzzy Bonferroni mean operator (HFFBM). Further, with the basic operations of HFFS in Einstein t-norms, the definition and derivation process of the hesitant Fermatean fuzzy Einstein Bonferroni mean operator (HFFEBM) are given. In addition, considering how weights affect decision-making outcomes, the hesitant Fermatean fuzzy weighted Bonferroni mean (HFFWBM) operator and the hesitant Fermatean fuzzy Einstein weighted Bonferroni mean operator (HFFEWBM) are developed. Then, the properties of the operators are discussed. Based on HFFWBM and HFFEWBM operator, a new multi-attribute decision-making (MADM) approach is provided. Finally, we apply the proposed decision-making approach to the case of a depression diagnostic evaluation for three depressed patients. The three patients' diagnosis results confirmed the proposed method's validity and rationality. Through a series of comparative experiments and analyses, the proposed MADM method is an efficient solution for decision-making issues in the hesitant Fermatean fuzzy environment.","PeriodicalId":10524,"journal":{"name":"Complex & Intelligent Systems","volume":"59 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-09-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"135879133","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：