NegCosIC: Negative Cosine Similarity-Invariance- Covariance Regularization for Few-Shot Learning
Few-shot learning continues to pose a challenge as it is inherently difficult for visual recognition models to generalize with limited labeled examples. When the training data is limited, the process of training and fine-tuning the model will be unstable and inefficient due to overfitting. In this p...
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
IEEE
2024-01-01
|
| Series: | IEEE Access |
| Subjects: | |
| Online Access: | https://ieeexplore.ieee.org/document/10495027/ |
| _version_ | 1797202222410366976 |
|---|---|
| author | Wei Han Liu Kian Ming Lim Thian Song Ong Chin Poo Lee |
| author_facet | Wei Han Liu Kian Ming Lim Thian Song Ong Chin Poo Lee |
| author_sort | Wei Han Liu |
| collection | DOAJ |
| description | Few-shot learning continues to pose a challenge as it is inherently difficult for visual recognition models to generalize with limited labeled examples. When the training data is limited, the process of training and fine-tuning the model will be unstable and inefficient due to overfitting. In this paper, we introduce NegCosIC: Negative Cosine Similarity-Invariance-Covariance Regularization, a method that aims to improve the mean accuracy from the perspective of stabilizing the fine-tuning process and regularizing variance. NegCosIC incorporates a negative simple cosine similarity loss to stabilize the parameters of the feature extractor during fine-tuning. In addition, NegCosIC integrates invariance loss and covariance loss to regularize the embeddings in order to reduce overfitting. Experimental results demonstrate that NegCosIC is able to bring substantial improvements over the current state-of-the-art methods. An in-depth worse case analysis is also conducted and shows that NegCosIC is able to outperform state-of-the-art methods on worst case accuracy. The proposed NegCosIC achieved 2.15% and 2.13% higher accuracy on miniImageNet 1-shot and 5-shot tasks, 3.22% and 2.67% higher accuracy on CUB 1-shot and 5-shot tasks, and 2.13% and 7.74% higher accuracy on CIFAR-FS 1-shot and 5-shot tasks in terms of worst-case accuracies. |
| first_indexed | 2024-04-24T08:00:00Z |
| format | Article |
| id | doaj.art-fcdc5b5adb714b9c9e1d3d9f195a18f6 |
| institution | Directory Open Access Journal |
| issn | 2169-3536 |
| language | English |
| last_indexed | 2024-04-24T08:00:00Z |
| publishDate | 2024-01-01 |
| publisher | IEEE |
| record_format | Article |
| series | IEEE Access |
| spelling | doaj.art-fcdc5b5adb714b9c9e1d3d9f195a18f62024-04-17T23:00:16ZengIEEEIEEE Access2169-35362024-01-0112528675287710.1109/ACCESS.2024.338680810495027NegCosIC: Negative Cosine Similarity-Invariance- Covariance Regularization for Few-Shot LearningWei Han Liu0https://orcid.org/0000-0002-4154-9572Kian Ming Lim1https://orcid.org/0000-0003-1929-7978Thian Song Ong2https://orcid.org/0000-0002-5867-9517Chin Poo Lee3https://orcid.org/0000-0003-3679-8977Faculty of Information Science and Technology, Multimedia University, Melaka, MalaysiaFaculty of Information Science and Technology, Multimedia University, Melaka, MalaysiaFaculty of Information Science and Technology, Multimedia University, Melaka, MalaysiaFaculty of Information Science and Technology, Multimedia University, Melaka, MalaysiaFew-shot learning continues to pose a challenge as it is inherently difficult for visual recognition models to generalize with limited labeled examples. When the training data is limited, the process of training and fine-tuning the model will be unstable and inefficient due to overfitting. In this paper, we introduce NegCosIC: Negative Cosine Similarity-Invariance-Covariance Regularization, a method that aims to improve the mean accuracy from the perspective of stabilizing the fine-tuning process and regularizing variance. NegCosIC incorporates a negative simple cosine similarity loss to stabilize the parameters of the feature extractor during fine-tuning. In addition, NegCosIC integrates invariance loss and covariance loss to regularize the embeddings in order to reduce overfitting. Experimental results demonstrate that NegCosIC is able to bring substantial improvements over the current state-of-the-art methods. An in-depth worse case analysis is also conducted and shows that NegCosIC is able to outperform state-of-the-art methods on worst case accuracy. The proposed NegCosIC achieved 2.15% and 2.13% higher accuracy on miniImageNet 1-shot and 5-shot tasks, 3.22% and 2.67% higher accuracy on CUB 1-shot and 5-shot tasks, and 2.13% and 7.74% higher accuracy on CIFAR-FS 1-shot and 5-shot tasks in terms of worst-case accuracies.https://ieeexplore.ieee.org/document/10495027/Few-shot learningnegative cosine similarityinvariancecovarianceregularization |
| spellingShingle | Wei Han Liu Kian Ming Lim Thian Song Ong Chin Poo Lee NegCosIC: Negative Cosine Similarity-Invariance- Covariance Regularization for Few-Shot Learning IEEE Access Few-shot learning negative cosine similarity invariance covariance regularization |
| title | NegCosIC: Negative Cosine Similarity-Invariance- Covariance Regularization for Few-Shot Learning |
| title_full | NegCosIC: Negative Cosine Similarity-Invariance- Covariance Regularization for Few-Shot Learning |
| title_fullStr | NegCosIC: Negative Cosine Similarity-Invariance- Covariance Regularization for Few-Shot Learning |
| title_full_unstemmed | NegCosIC: Negative Cosine Similarity-Invariance- Covariance Regularization for Few-Shot Learning |
| title_short | NegCosIC: Negative Cosine Similarity-Invariance- Covariance Regularization for Few-Shot Learning |
| title_sort | negcosic negative cosine similarity invariance covariance regularization for few shot learning |
| topic | Few-shot learning negative cosine similarity invariance covariance regularization |
| url | https://ieeexplore.ieee.org/document/10495027/ |
| work_keys_str_mv | AT weihanliu negcosicnegativecosinesimilarityinvariancecovarianceregularizationforfewshotlearning AT kianminglim negcosicnegativecosinesimilarityinvariancecovarianceregularizationforfewshotlearning AT thiansongong negcosicnegativecosinesimilarityinvariancecovarianceregularizationforfewshotlearning AT chinpoolee negcosicnegativecosinesimilarityinvariancecovarianceregularizationforfewshotlearning |