Pub Date : 2023-03-01DOI: 10.1109/icnlp58431.2023.00028
Jin Wu, Zhengdong Su, L. Yang, Yaqioang Gao
In recent years, with the development of Internet of Things (IoT) technology, many things around us are closely connected through network communication. In this paper, a Cyclic Shift Chirp Spread Spectrum (CS-CSS) technique is proposed by combining the reference coefficients and design indexes defined by the Long Range (LoRa) protocol. The technology combines Chirp spread spectrum (CSS) with Cyclic Code Shift Keying (CCSK) coding spread spectrum, maps the input data on Cyclic Shift Factor (CSF), and solves the corresponding Cyclic Shift Factor by FFT at the receiver. Compared with Chirp-BOK system, it has better Bit Error Rate(BER) performance and stronger anti-interference ability. In terms of performance, compared with the parameters and indicators specified in LoRaIoT protocol, it can meet the requirements. Then, the experiment shows that compared with the Chirp-BOK system, the BER performance has more than 10dB gain when the Spread Factor (SF) is 7, and the modulation efficiency of the system wm also increase or decrease with the change. Finally, the synchronization scheme is studied and the algorithms for estimating time offset and frequency offset are designed. The experimental results show that the proposed algorithm has good performance. In this paper, some advantages and contributions based on the proposed technology are also described in view of many bottlenecks facing the IoT industry at the present stage, such as cost, power consumption, synchronization implementation, modulation efficiency and other issues.
{"title":"Research on CS-CSS Modulation System Based on Chirp Signal","authors":"Jin Wu, Zhengdong Su, L. Yang, Yaqioang Gao","doi":"10.1109/icnlp58431.2023.00028","DOIUrl":"https://doi.org/10.1109/icnlp58431.2023.00028","url":null,"abstract":"In recent years, with the development of Internet of Things (IoT) technology, many things around us are closely connected through network communication. In this paper, a Cyclic Shift Chirp Spread Spectrum (CS-CSS) technique is proposed by combining the reference coefficients and design indexes defined by the Long Range (LoRa) protocol. The technology combines Chirp spread spectrum (CSS) with Cyclic Code Shift Keying (CCSK) coding spread spectrum, maps the input data on Cyclic Shift Factor (CSF), and solves the corresponding Cyclic Shift Factor by FFT at the receiver. Compared with Chirp-BOK system, it has better Bit Error Rate(BER) performance and stronger anti-interference ability. In terms of performance, compared with the parameters and indicators specified in LoRaIoT protocol, it can meet the requirements. Then, the experiment shows that compared with the Chirp-BOK system, the BER performance has more than 10dB gain when the Spread Factor (SF) is 7, and the modulation efficiency of the system wm also increase or decrease with the change. Finally, the synchronization scheme is studied and the algorithms for estimating time offset and frequency offset are designed. The experimental results show that the proposed algorithm has good performance. In this paper, some advantages and contributions based on the proposed technology are also described in view of many bottlenecks facing the IoT industry at the present stage, such as cost, power consumption, synchronization implementation, modulation efficiency and other issues.","PeriodicalId":53637,"journal":{"name":"Icon","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2023-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"82532905","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 : 2023-03-01DOI: 10.1109/ICNLP58431.2023.00051
Maurilio De Araujo Possi, Alcione de Paiva Oliveira, Alexandra Moreira, Lucas Mucida Costa
Automatic poem generation has been a challenging topic in Natural Language Processing research. However, their evaluation is still largely based on methods that involve evaluation by human judges, comparing the results with poems written by humans, or using metrics that were not designed for this purpose. In order to fill this gap, this work proposes a specific metric for automatic evaluation of poems, capable of quantitatively evaluating morphological characteristics of the text, such as rhyme and meter, of different types within this literary genre. The tests carried out demonstrate that the metric developed presents an advance for the evaluation of this type of text.
{"title":"CARMEN: A Method for Automatic Evaluation of Poems","authors":"Maurilio De Araujo Possi, Alcione de Paiva Oliveira, Alexandra Moreira, Lucas Mucida Costa","doi":"10.1109/ICNLP58431.2023.00051","DOIUrl":"https://doi.org/10.1109/ICNLP58431.2023.00051","url":null,"abstract":"Automatic poem generation has been a challenging topic in Natural Language Processing research. However, their evaluation is still largely based on methods that involve evaluation by human judges, comparing the results with poems written by humans, or using metrics that were not designed for this purpose. In order to fill this gap, this work proposes a specific metric for automatic evaluation of poems, capable of quantitatively evaluating morphological characteristics of the text, such as rhyme and meter, of different types within this literary genre. The tests carried out demonstrate that the metric developed presents an advance for the evaluation of this type of text.","PeriodicalId":53637,"journal":{"name":"Icon","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2023-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"85574779","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 : 2023-03-01DOI: 10.1109/ICNLP58431.2023.00042
Min Zhang, Yanqing Zhao, Xiaosong Qiao, Song Peng, Shimin Tao, Hao Yang, Ying Qin, Yanfei Jiang
Named Entity Recognition (NER) is one of the most fundamental tasks in natural language processing (NLP). Different from the widely-used sequence labeling framework in NER, span prediction based methods are more naturally suitable for the nested NER problem and have received a lot of attention recently. However, classifying the samples generated by traversing all sub-sequences is computational expensive during training and very ineffective at inference. In this paper, we propose the FastSpanNER approach to reduce the computation of both training and inferring. We introduce a task of Named Entity Head (NEH) prediction for each word in given sequence, and perform multi-task learning together with the task of span classification, which uses no more than half of the samples in SpanNER. In the inference phase, only the words predicted as NEHs are used to generate candidate spans for named entity classification. Experimental results on the four standard benchmark datasets (CoNLL2003, MSRA, CNERTA and GENIA) show that our FastSpanNER method not only greatly reduces the computation of training and inferring but also achieves better F1 scores compared with the SpanNER method.
{"title":"FastSpanNER: Speeding up SpanNER by Named Entity Head Prediction","authors":"Min Zhang, Yanqing Zhao, Xiaosong Qiao, Song Peng, Shimin Tao, Hao Yang, Ying Qin, Yanfei Jiang","doi":"10.1109/ICNLP58431.2023.00042","DOIUrl":"https://doi.org/10.1109/ICNLP58431.2023.00042","url":null,"abstract":"Named Entity Recognition (NER) is one of the most fundamental tasks in natural language processing (NLP). Different from the widely-used sequence labeling framework in NER, span prediction based methods are more naturally suitable for the nested NER problem and have received a lot of attention recently. However, classifying the samples generated by traversing all sub-sequences is computational expensive during training and very ineffective at inference. In this paper, we propose the FastSpanNER approach to reduce the computation of both training and inferring. We introduce a task of Named Entity Head (NEH) prediction for each word in given sequence, and perform multi-task learning together with the task of span classification, which uses no more than half of the samples in SpanNER. In the inference phase, only the words predicted as NEHs are used to generate candidate spans for named entity classification. Experimental results on the four standard benchmark datasets (CoNLL2003, MSRA, CNERTA and GENIA) show that our FastSpanNER method not only greatly reduces the computation of training and inferring but also achieves better F1 scores compared with the SpanNER method.","PeriodicalId":53637,"journal":{"name":"Icon","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2023-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"80076878","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 : 2023-03-01DOI: 10.1109/ICNLP58431.2023.00045
Omar Alsayed, Chloe Muncy, Ahmed Youssef, Ryan Green
Unsupervised text style transfer methods aim to transfer the style of the text without affecting its fundamental meaning using non-parallel data. Although previous work has explored few-shot learning for this task, incorporating few-shot abstractive summarization and its benefits have not yet been explored. Hence, we propose a novel unsupervised text style transfer approach using few-shot abstractive summarization. In our method, we infer a vector space embedding for the corpora and align the source-target embeddings using their vector space centroids. A set of nearest neighbors is retrieved for every source text unit from the target style based on their semantic similarity in the aligned vector space. Multiple subsets of nearest neighbors are extracted and summarized using a language model with a reranking procedure to optimize the style transfer quality, which achieves state-of-the-art results on automatic evaluation metrics.
{"title":"FASST: Few-Shot Abstractive Summarization for Style Transfer","authors":"Omar Alsayed, Chloe Muncy, Ahmed Youssef, Ryan Green","doi":"10.1109/ICNLP58431.2023.00045","DOIUrl":"https://doi.org/10.1109/ICNLP58431.2023.00045","url":null,"abstract":"Unsupervised text style transfer methods aim to transfer the style of the text without affecting its fundamental meaning using non-parallel data. Although previous work has explored few-shot learning for this task, incorporating few-shot abstractive summarization and its benefits have not yet been explored. Hence, we propose a novel unsupervised text style transfer approach using few-shot abstractive summarization. In our method, we infer a vector space embedding for the corpora and align the source-target embeddings using their vector space centroids. A set of nearest neighbors is retrieved for every source text unit from the target style based on their semantic similarity in the aligned vector space. Multiple subsets of nearest neighbors are extracted and summarized using a language model with a reranking procedure to optimize the style transfer quality, which achieves state-of-the-art results on automatic evaluation metrics.","PeriodicalId":53637,"journal":{"name":"Icon","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2023-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"81471939","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 : 2023-03-01DOI: 10.1109/ICNLP58431.2023.00032
Xi Hai Xie, Sheng Yuan Na
Coherent accumulation is commonly used to improve the detection capability of radar systems in cluttered environments. Based on the principle of moving target detection, two algorithms for implementing coherent accumulation are proposed in this paper. After comparing their complexity and computational effort, an optimal method is chosen to implement coherent accumulation based on the Keystone transform.
{"title":"Long-term Coherent Accumulation Algorithm Based on Radar Altimeter","authors":"Xi Hai Xie, Sheng Yuan Na","doi":"10.1109/ICNLP58431.2023.00032","DOIUrl":"https://doi.org/10.1109/ICNLP58431.2023.00032","url":null,"abstract":"Coherent accumulation is commonly used to improve the detection capability of radar systems in cluttered environments. Based on the principle of moving target detection, two algorithms for implementing coherent accumulation are proposed in this paper. After comparing their complexity and computational effort, an optimal method is chosen to implement coherent accumulation based on the Keystone transform.","PeriodicalId":53637,"journal":{"name":"Icon","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2023-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"72680744","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 : 2023-03-01DOI: 10.1109/ICNLP58431.2023.00052
Jihad R’Baiti, R. Faizi, Youssef Hmamouche, A. E. Seghrouchni
As technology advances, everything is becoming digitized, including newspapers and magazines. Currently, information is accessible in an easy, and fast manner. However, some content creators exploit this opportunity negatively by using unethical methods to attract users’ attention aiming to increase their ads’ income instead of providing accurate information. In this research, we propose a comparative study of various approaches based on natural language processing techniques and deep learning models to face this clickbait phenomenon. This study will enable us to detect this type of content in Arabic. Fine-tuned BERT with an attached neural network layer architecture achieved the highest results with an accuracy of 0.9103, a precision of 0.9111, and a recall of 0.9103 outperformed CNN, LSTM, BiLSTM, and FFNN using the different representation methods TF-IDF, Roberta, and Embedding.
{"title":"A transformer-based architecture for the automatic detection of clickbait for Arabic headlines","authors":"Jihad R’Baiti, R. Faizi, Youssef Hmamouche, A. E. Seghrouchni","doi":"10.1109/ICNLP58431.2023.00052","DOIUrl":"https://doi.org/10.1109/ICNLP58431.2023.00052","url":null,"abstract":"As technology advances, everything is becoming digitized, including newspapers and magazines. Currently, information is accessible in an easy, and fast manner. However, some content creators exploit this opportunity negatively by using unethical methods to attract users’ attention aiming to increase their ads’ income instead of providing accurate information. In this research, we propose a comparative study of various approaches based on natural language processing techniques and deep learning models to face this clickbait phenomenon. This study will enable us to detect this type of content in Arabic. Fine-tuned BERT with an attached neural network layer architecture achieved the highest results with an accuracy of 0.9103, a precision of 0.9111, and a recall of 0.9103 outperformed CNN, LSTM, BiLSTM, and FFNN using the different representation methods TF-IDF, Roberta, and Embedding.","PeriodicalId":53637,"journal":{"name":"Icon","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2023-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"78159932","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 : 2023-03-01DOI: 10.1109/ICNLP58431.2023.00038
Qishun Mei
Sentiment classification is a classical and important task of natural language processing (NLP), with the development of the Internet, there are multifarious reviews, comments and news produced everyday which need high cost to annotate, so it has become a challenge to develop a more effective sentiment classification model which requires less training samples. In this paper, we propose a sentence level sentiment classification model based on Contrastive Learning, Generative Adversarial Network and BERT (CON-GAN-BERT). Experiments on several public Chinese sentiment classification datasets show that CON-GAN-BERT significantly outperforms strong pre-training baseline, and still obtaining good performances for Few-Shot Learning without any data augmentation or unlabeled data.
{"title":"CON-GAN-BERT: combining Contrastive Learning with Generative Adversarial Nets for Few-Shot Sentiment Classification","authors":"Qishun Mei","doi":"10.1109/ICNLP58431.2023.00038","DOIUrl":"https://doi.org/10.1109/ICNLP58431.2023.00038","url":null,"abstract":"Sentiment classification is a classical and important task of natural language processing (NLP), with the development of the Internet, there are multifarious reviews, comments and news produced everyday which need high cost to annotate, so it has become a challenge to develop a more effective sentiment classification model which requires less training samples. In this paper, we propose a sentence level sentiment classification model based on Contrastive Learning, Generative Adversarial Network and BERT (CON-GAN-BERT). Experiments on several public Chinese sentiment classification datasets show that CON-GAN-BERT significantly outperforms strong pre-training baseline, and still obtaining good performances for Few-Shot Learning without any data augmentation or unlabeled data.","PeriodicalId":53637,"journal":{"name":"Icon","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2023-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"73695335","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}
Vehicle-to-Everything (V2X) communication is a groundbreaking technology that enables interconnected services in the realm of smart transportation. Among the various V2X applications, Vehicle-to-Pedestrian (V2P) communication plays a crucial role in enhancing road traffic efficiency and safety by facilitating the exchange of information between vehicles and pedestrians. However, the existing V2P warning systems neglect the inherent uncertainty associated with pedestrian trajectories, leading to suboptimal accuracy in detecting collision risks between vehicles and pedestrians. Consequently, the potential for improving road safety is limited. To address this issue, we propose an advanced pedestrian-vehicle anti-collision model. This model takes into account the uncertain nature of pedestrian movement and leverages the Long Range (LoRa) wireless network to establish a V2P warning system. Specifically, we employ the long short-term memory artificial neural network (LSTM) to accurately predict pedestrian trajectories. By combining the pedestrian’s trajectory with a multi-dimensional normal distribution function, we obtain the probability density function that characterizes the pedestrian’s movement. Subsequently, we deduce the preliminary collision area between pedestrians and vehicles. Finally, we utilize a confidence probability metric to determine whether a warning should be issued to both pedestrians and vehicles. Simulation results demonstrate the effectiveness of our system in accurately warning pedestrians and vehicles, even under varying speeds and Global Positioning System (GPS) positioning errors. The experimental evaluation of our proposed method further validates its superior performance and efficacy.
{"title":"A V2P Warning System on the Basis of LoRa Wireless Network","authors":"Ruoyu Pan, Lihua Jie, Honggang Wang, Peihua Jie, Xinyue Zhang","doi":"10.1109/ICNLP58431.2023.00074","DOIUrl":"https://doi.org/10.1109/ICNLP58431.2023.00074","url":null,"abstract":"Vehicle-to-Everything (V2X) communication is a groundbreaking technology that enables interconnected services in the realm of smart transportation. Among the various V2X applications, Vehicle-to-Pedestrian (V2P) communication plays a crucial role in enhancing road traffic efficiency and safety by facilitating the exchange of information between vehicles and pedestrians. However, the existing V2P warning systems neglect the inherent uncertainty associated with pedestrian trajectories, leading to suboptimal accuracy in detecting collision risks between vehicles and pedestrians. Consequently, the potential for improving road safety is limited. To address this issue, we propose an advanced pedestrian-vehicle anti-collision model. This model takes into account the uncertain nature of pedestrian movement and leverages the Long Range (LoRa) wireless network to establish a V2P warning system. Specifically, we employ the long short-term memory artificial neural network (LSTM) to accurately predict pedestrian trajectories. By combining the pedestrian’s trajectory with a multi-dimensional normal distribution function, we obtain the probability density function that characterizes the pedestrian’s movement. Subsequently, we deduce the preliminary collision area between pedestrians and vehicles. Finally, we utilize a confidence probability metric to determine whether a warning should be issued to both pedestrians and vehicles. Simulation results demonstrate the effectiveness of our system in accurately warning pedestrians and vehicles, even under varying speeds and Global Positioning System (GPS) positioning errors. The experimental evaluation of our proposed method further validates its superior performance and efficacy.","PeriodicalId":53637,"journal":{"name":"Icon","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2023-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"73989095","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 : 2023-03-01DOI: 10.1109/ICNLP58431.2023.00057
Nikhil Yelamarthy, Oshin Anand
This paper primarily focuses on developing an end-to-end solution which can process multi-threaded conversations and perform information extraction (IE) specific to a domain and intended business task. The challenges of IE in a conversation are a) context understanding, which consists of two elements: topic and sense of expression and b) establishing context flow. Since the target is free-flow dialogue, understanding the change in contexts is crucial. In this research, we attempt to build a solution that can infer and connect these contexts and reflect the same in the extracted information, taking care of things like negotiations. The proposed approach has three main steps; The first step is domain-dependent which performs topic classification at the sentence level. The second step is domain-independent, and it categorizes sentences into different semantic classes, to understand the conversation flow and parse it into multiple conversation threads. In the final step, we carry out morphological parsing to extract the target value, utilizing the predicted sentence class labels along with the conversation flow. A buyer-seller chat conversation is taken as the sample domain and the target IE is towards information for purchase order generation.
{"title":"Context-aware Information Extraction from Multi-thread Business Conversations","authors":"Nikhil Yelamarthy, Oshin Anand","doi":"10.1109/ICNLP58431.2023.00057","DOIUrl":"https://doi.org/10.1109/ICNLP58431.2023.00057","url":null,"abstract":"This paper primarily focuses on developing an end-to-end solution which can process multi-threaded conversations and perform information extraction (IE) specific to a domain and intended business task. The challenges of IE in a conversation are a) context understanding, which consists of two elements: topic and sense of expression and b) establishing context flow. Since the target is free-flow dialogue, understanding the change in contexts is crucial. In this research, we attempt to build a solution that can infer and connect these contexts and reflect the same in the extracted information, taking care of things like negotiations. The proposed approach has three main steps; The first step is domain-dependent which performs topic classification at the sentence level. The second step is domain-independent, and it categorizes sentences into different semantic classes, to understand the conversation flow and parse it into multiple conversation threads. In the final step, we carry out morphological parsing to extract the target value, utilizing the predicted sentence class labels along with the conversation flow. A buyer-seller chat conversation is taken as the sample domain and the target IE is towards information for purchase order generation.","PeriodicalId":53637,"journal":{"name":"Icon","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2023-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"74265278","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：