Search Results for author:
Found 3 papers, 1 papers with code
Borrowing Treasures from Neighbors: In-Context Learning for Multimodal Learning with Missing Modalities and Data Scarcity
The proposed data-dependent framework exhibits a higher degree of sample efficiency and is empirically demonstrated to enhance the classification model's performance on both full- and missing-modality data in the low-data regime across various multimodal learning tasks.
HgbNet: predicting hemoglobin level/anemia degree from EHR data
EHR-based hemoglobin level/anemia degree prediction is non-invasive and rapid but still faces some challenges due to the fact that EHR data is typically an irregular multivariate time series containing a significant number of missing values and irregular time intervals.
Efficient Approximation of Action Potentials with High-Order Shape Preservation in Unsupervised Spike Sorting
The use of the Taylor polynomial is proposed and modelled employing its cascaded derivatives to non-uniformly capture the essential samples in each spike for reliable feature extraction and sorting.
Cannot find the paper you are looking for? You can
Submit a new open access paper.
