Open-Set Signal Recognition Based on Transformer and Wasserstein Distance
Open-set signal recognition provides a new approach for verifying the robustness of models by introducing novel unknown signal classes into the model testing and breaking the conventional closed-set assumption, which has become very popular in real-world scenarios. In the present work, we propose an...
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2023-02-01
|
| Series: | Applied Sciences |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2076-3417/13/4/2151 |
| _version_ | 1797622609530060800 |
|---|---|
| author | Wei Zhang Da Huang Minghui Zhou Jingran Lin Xiangfeng Wang |
| author_facet | Wei Zhang Da Huang Minghui Zhou Jingran Lin Xiangfeng Wang |
| author_sort | Wei Zhang |
| collection | DOAJ |
| description | Open-set signal recognition provides a new approach for verifying the robustness of models by introducing novel unknown signal classes into the model testing and breaking the conventional closed-set assumption, which has become very popular in real-world scenarios. In the present work, we propose an efficient open-set signal recognition algorithm, which contains three key sub-modules: the signal representation sub-module based on a vision transformer (ViT) structure, a set distance metric sub-module based on Wasserstein distance, and a class space compression sub-module based on reciprocal point separation and central loss. In this algorithm, the representing features of signals are established based on transformer-based neural networks, i.e., ViT, in order to extract global information about time series-related data. The employed reciprocal point is used in modeling the potential unknown space without using the corresponding samples, while the distance metric between different class spaces is mathematically modeled in terms of the Wasserstein distance instead of the classical Euclidean distance. Numerical experiments on different open-set signal recognition tasks show that the proposed algorithm can significantly improve the recognition efficiency in both known and unknown categories. |
| first_indexed | 2024-03-11T09:12:45Z |
| format | Article |
| id | doaj.art-2aca3b89ec434308b7732777d5c9ab59 |
| institution | Directory Open Access Journal |
| issn | 2076-3417 |
| language | English |
| last_indexed | 2024-03-11T09:12:45Z |
| publishDate | 2023-02-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Applied Sciences |
| spelling | doaj.art-2aca3b89ec434308b7732777d5c9ab592023-11-16T18:51:36ZengMDPI AGApplied Sciences2076-34172023-02-01134215110.3390/app13042151Open-Set Signal Recognition Based on Transformer and Wasserstein DistanceWei Zhang0Da Huang1Minghui Zhou2Jingran Lin3Xiangfeng Wang4School of Information and Communication Engineering, University of Electronic Science and Technology of China, Chengdu 611731, ChinaSchool of Computer Science and Technology, East China Normal University, Shanghai 200062, ChinaSchool of Computer Science and Technology, East China Normal University, Shanghai 200062, ChinaSchool of Information and Communication Engineering, University of Electronic Science and Technology of China, Chengdu 611731, ChinaSchool of Computer Science and Technology, East China Normal University, Shanghai 200062, ChinaOpen-set signal recognition provides a new approach for verifying the robustness of models by introducing novel unknown signal classes into the model testing and breaking the conventional closed-set assumption, which has become very popular in real-world scenarios. In the present work, we propose an efficient open-set signal recognition algorithm, which contains three key sub-modules: the signal representation sub-module based on a vision transformer (ViT) structure, a set distance metric sub-module based on Wasserstein distance, and a class space compression sub-module based on reciprocal point separation and central loss. In this algorithm, the representing features of signals are established based on transformer-based neural networks, i.e., ViT, in order to extract global information about time series-related data. The employed reciprocal point is used in modeling the potential unknown space without using the corresponding samples, while the distance metric between different class spaces is mathematically modeled in terms of the Wasserstein distance instead of the classical Euclidean distance. Numerical experiments on different open-set signal recognition tasks show that the proposed algorithm can significantly improve the recognition efficiency in both known and unknown categories.https://www.mdpi.com/2076-3417/13/4/2151open-set classificationsignal recognitiontransformerViTWasserstein distance |
| spellingShingle | Wei Zhang Da Huang Minghui Zhou Jingran Lin Xiangfeng Wang Open-Set Signal Recognition Based on Transformer and Wasserstein Distance Applied Sciences open-set classification signal recognition transformer ViT Wasserstein distance |
| title | Open-Set Signal Recognition Based on Transformer and Wasserstein Distance |
| title_full | Open-Set Signal Recognition Based on Transformer and Wasserstein Distance |
| title_fullStr | Open-Set Signal Recognition Based on Transformer and Wasserstein Distance |
| title_full_unstemmed | Open-Set Signal Recognition Based on Transformer and Wasserstein Distance |
| title_short | Open-Set Signal Recognition Based on Transformer and Wasserstein Distance |
| title_sort | open set signal recognition based on transformer and wasserstein distance |
| topic | open-set classification signal recognition transformer ViT Wasserstein distance |
| url | https://www.mdpi.com/2076-3417/13/4/2151 |
| work_keys_str_mv | AT weizhang opensetsignalrecognitionbasedontransformerandwassersteindistance AT dahuang opensetsignalrecognitionbasedontransformerandwassersteindistance AT minghuizhou opensetsignalrecognitionbasedontransformerandwassersteindistance AT jingranlin opensetsignalrecognitionbasedontransformerandwassersteindistance AT xiangfengwang opensetsignalrecognitionbasedontransformerandwassersteindistance |