A Novel Ensemble Strategy Based on Determinantal Point Processes for Transfer Learning
Transfer learning (TL) hopes to train a model for target domain tasks by using knowledge from different but related source domains. Most TL methods focus more on improving the predictive performance of the single model across domains. Since domain differences cannot be avoided, the knowledge from th...
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2022-11-01
|
| Series: | Mathematics |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2227-7390/10/23/4409 |
| _version_ | 1827643008477036544 |
|---|---|
| author | Ying Lv Bofeng Zhang Xiaodong Yue Zhikang Xu |
| author_facet | Ying Lv Bofeng Zhang Xiaodong Yue Zhikang Xu |
| author_sort | Ying Lv |
| collection | DOAJ |
| description | Transfer learning (TL) hopes to train a model for target domain tasks by using knowledge from different but related source domains. Most TL methods focus more on improving the predictive performance of the single model across domains. Since domain differences cannot be avoided, the knowledge from the source domain to obtain the target domain is limited. Therefore, the transfer model has to predict out-of-distribution (OOD) data in the target domain. However, the prediction of the single model is unstable when dealing with the OOD data, which can easily cause negative transfer. To solve this problem, we propose a parallel ensemble strategy based on Determinantal Point Processes (DPP) for transfer learning. In this strategy, we first proposed an improved DPP sampling to generate training subsets with higher transferability and diversity. Second, we use the subsets to train the base models. Finally, the base models are fused using the adaptability of subsets. To validate the effectiveness of the ensemble strategy, we couple the ensemble strategy into traditional TL models and deep TL models and evaluate the transfer performance models on text and image data sets. The experiment results show that our proposed ensemble strategy can significantly improve the performance of the transfer model. |
| first_indexed | 2024-03-09T17:40:46Z |
| format | Article |
| id | doaj.art-7cc9cad8c8004d5e93024946f088f5ac |
| institution | Directory Open Access Journal |
| issn | 2227-7390 |
| language | English |
| last_indexed | 2024-03-09T17:40:46Z |
| publishDate | 2022-11-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Mathematics |
| spelling | doaj.art-7cc9cad8c8004d5e93024946f088f5ac2023-11-24T11:33:04ZengMDPI AGMathematics2227-73902022-11-011023440910.3390/math10234409A Novel Ensemble Strategy Based on Determinantal Point Processes for Transfer LearningYing Lv0Bofeng Zhang1Xiaodong Yue2Zhikang Xu3School of Computer Engineering and Science, Shanghai University, Shanghai 200444, ChinaSchool of Computer and Information Engineering, Shanghai Polytechnic University, Shanghai 201209, ChinaSchool of Computer Engineering and Science, Shanghai University, Shanghai 200444, ChinaSchool of Computer Engineering and Science, Shanghai University, Shanghai 200444, ChinaTransfer learning (TL) hopes to train a model for target domain tasks by using knowledge from different but related source domains. Most TL methods focus more on improving the predictive performance of the single model across domains. Since domain differences cannot be avoided, the knowledge from the source domain to obtain the target domain is limited. Therefore, the transfer model has to predict out-of-distribution (OOD) data in the target domain. However, the prediction of the single model is unstable when dealing with the OOD data, which can easily cause negative transfer. To solve this problem, we propose a parallel ensemble strategy based on Determinantal Point Processes (DPP) for transfer learning. In this strategy, we first proposed an improved DPP sampling to generate training subsets with higher transferability and diversity. Second, we use the subsets to train the base models. Finally, the base models are fused using the adaptability of subsets. To validate the effectiveness of the ensemble strategy, we couple the ensemble strategy into traditional TL models and deep TL models and evaluate the transfer performance models on text and image data sets. The experiment results show that our proposed ensemble strategy can significantly improve the performance of the transfer model.https://www.mdpi.com/2227-7390/10/23/4409transfer learningensemble strategydeterminantal point processesdomain adaptation |
| spellingShingle | Ying Lv Bofeng Zhang Xiaodong Yue Zhikang Xu A Novel Ensemble Strategy Based on Determinantal Point Processes for Transfer Learning Mathematics transfer learning ensemble strategy determinantal point processes domain adaptation |
| title | A Novel Ensemble Strategy Based on Determinantal Point Processes for Transfer Learning |
| title_full | A Novel Ensemble Strategy Based on Determinantal Point Processes for Transfer Learning |
| title_fullStr | A Novel Ensemble Strategy Based on Determinantal Point Processes for Transfer Learning |
| title_full_unstemmed | A Novel Ensemble Strategy Based on Determinantal Point Processes for Transfer Learning |
| title_short | A Novel Ensemble Strategy Based on Determinantal Point Processes for Transfer Learning |
| title_sort | novel ensemble strategy based on determinantal point processes for transfer learning |
| topic | transfer learning ensemble strategy determinantal point processes domain adaptation |
| url | https://www.mdpi.com/2227-7390/10/23/4409 |
| work_keys_str_mv | AT yinglv anovelensemblestrategybasedondeterminantalpointprocessesfortransferlearning AT bofengzhang anovelensemblestrategybasedondeterminantalpointprocessesfortransferlearning AT xiaodongyue anovelensemblestrategybasedondeterminantalpointprocessesfortransferlearning AT zhikangxu anovelensemblestrategybasedondeterminantalpointprocessesfortransferlearning AT yinglv novelensemblestrategybasedondeterminantalpointprocessesfortransferlearning AT bofengzhang novelensemblestrategybasedondeterminantalpointprocessesfortransferlearning AT xiaodongyue novelensemblestrategybasedondeterminantalpointprocessesfortransferlearning AT zhikangxu novelensemblestrategybasedondeterminantalpointprocessesfortransferlearning |