Abstract
Deep transfer learning aims at dealing with challenges in new tasks with insufficient samples. However, when it comes to few-shot learning scenarios, due to the low diversity of several known training samples, they are prone to be dominated by specificity, thus leading to one-sidedness local features instead of the reliable global feature of the actual categories they belong to. To alleviate the difficulty, we propose a cross-modal few-shot contextual transfer method that leverages the contextual information as a supplement and learns context awareness transfer in few-shot image classification scenes, which fully utilizes the information in heterogeneous data. The similarity measure in the image classification task is reformulated via fusing textual semantic modal information and visual semantic modal information extracted from images. This performs as a supplement and helps to inhibit the sample specificity. Besides, to better extract local visual features and reorganize the recognition pattern, the deep transfer scheme is also used for reusing a powerful extractor from the pre-trained model. Simulation experiments show that the introduction of cross-modal and intra-modal contextual information can effectively suppress the deviation of defining category features with few samples and improve the accuracy of few-shot image classification tasks.
Highlights
In the age of big data, data are vast but sometimes precious
Classical convolutional neural network (CNN)-based image classification tasks often fail in directly using a few-shot learning scenario, since the models rely on well-trained feature extractors and classifiers, which need a large number of samples to adequately learn interior features
Concerning the architecture of the training model with the proposed contextual transfer scheme, we choose the same network structures as each previous method we used for comparison, because the convolutional layers for feature learning can be flexibly adjusted in the proposed architecture, such as the designed CNN net with four convolutional modules and the ResNet-12, while the deep transfer learning (DTL) scheme is introduced for adaptation
Summary
Deep neutral networks have achieved great performance when sufficient samples are offered. When it comes to practical tasks, these frameworks fail due to small data limitations. Few-shot learning is proposed to simulate a human learner who is adept at generalizing categories with a handful of samples, and it fits well with a possible lack of data in the task of actual solutions (Lu et al, 2020), leading to an innovation of artificial intelligence algorithms. Since deep transfer learning (DTL) has been a good solution in actual tasks to transfer and reuse knowledge from the auxiliary domain, researchers have been exploring whether some important information can
Sign-in/Register to view highlights & summary Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
