Attention and feature transfer based knowledge distillation [PDF]
Scientific Reports, 2023 Existing knowledge distillation (KD) methods are mainly based on features, logic, or attention, where features and logic represent the results of reasoning at different stages of a convolutional neural network, and attention maps symbolize the reasoning ...
Guoliang Yang +3 more
doaj +4 more sources
Attention-based speech feature transfer between speakers [PDF]
Frontiers in Artificial IntelligenceIn this study, we propose a simple yet effective method for incorporating the source speaker's characteristics in the target speaker's speech. This allows our model to generate the speech of the target speaker with the style of the source speaker.
Hangbok Lee, Minjae Cho, Hyuk-Yoon Kwon
doaj +4 more sources
Semi-Supervised Instance-Segmentation Model for Feature Transfer Based on Category Attention [PDF]
Sensors, 2022 In the task of image instance segmentation, semi-supervised instance segmentation algorithms have received constant research attention over recent years.
Hao Wang +7 more
doaj +2 more sources
Joint Feature-Space and Sample-Space Based Heterogeneous Feature Transfer Method for Object Recognition Using Remote Sensing Images with Different Spatial Resolutions [PDF]
Sensors, 2021 To improve the classification results of high-resolution remote sensing images (RSIs), it is necessary to use feature transfer methods to mine the relevant information between high-resolution RSIs and low-resolution RSIs to train the classifiers together.
Wei Hu +7 more
doaj +2 more sources
Attention-based Wav2Text with Feature Transfer Learning [PDF]
2017 IEEE Automatic Speech Recognition and Understanding Workshop (ASRU), 2017 Conventional automatic speech recognition (ASR) typically performs multi-level pattern recognition tasks that map the acoustic speech waveform into a hierarchy of speech units.
Nakamura, Satoshi +2 more
core +2 more sources
Transfer Learning Based Data Feature Transfer for Fault Diagnosis [PDF]
IEEE Access, 2020 The development of sensor technology provides massive data for data-driven fault diagnosis. In recent years, more and more scholars are studying artificial intelligence technology to solve the bottleneck in fault diagnosis.
Wei Xu, Yi Wan, Tian-Yu Zuo, Xin-Mei Sha
doaj +2 more sources
Stimulus blanking reveals contrast-dependent transsaccadic feature transfer. [PDF]
Sci Rep, 2020 Abstract Across saccadic eye movements, the visual system receives two successive static images corresponding to the pre- and the postsaccadic projections of the visual field on the retina. The existence of a mechanism integrating the content of these images is today still a matter of debate.
Grzeczkowski L, Deubel H, Szinte M.
europepmc +5 more sources
Partial Transfer Learning Method Based on Inter-Class Feature Transfer for Rolling Bearing Fault Diagnosis [PDF]
SensorsRolling bearing fault diagnosis methods based on transfer learning always assume that the sample classes in the target domain are consistent with those in the source domain during the training phase.
Hongbo Que +6 more
doaj +2 more sources
The Transmission of Substrate Features
Journal of Modern Languages, 2020 Theories of the formation of creole syntax have been proposed to explain whether substrates and superstrates influence the resultant creole structures, and if so, what the mechanisms are by which they influence them.
Robert Laub
doaj +14 more sources
Feature-Supervised Action Modality Transfer [PDF]
2020 25th International Conference on Pattern Recognition (ICPR), 2021 This paper strives for action recognition and detection in video modalities like RGB, depth maps or 3D-skeleton sequences when only limited modality-specific labeled examples are available. For the RGB, and derived optical-flow, modality many large-scale labeled datasets have been made available.
Thoker, F.M., Snoek, C.G.M.
openaire +4 more sources