Pmr: Prototypical modal rebalance for multimodal learning
Multimodal learning (MML) aims to jointly exploit the common priors of different modalities to
compensate for their inherent limitations. However, existing MML methods often optimize a …
compensate for their inherent limitations. However, existing MML methods often optimize a …
Prototype-based sample-weighted distillation unified framework adapted to missing modality sentiment analysis
Y Zhang, X Zhuang, Y Hou, Y Zhang - Neural Networks, 2024 - Elsevier
Missing modality sentiment analysis is a prevalent and challenging issue in real life.
Furthermore, the heterogeneity of multimodality often leads to an imbalance in optimization …
Furthermore, the heterogeneity of multimodality often leads to an imbalance in optimization …
Few-shot electronic health record coding through graph contrastive learning
Electronic health record (EHR) coding is the task of assigning ICD codes to each EHR. Most
previous studies either only focus on the frequent ICD codes or treat rare and frequent ICD …
previous studies either only focus on the frequent ICD codes or treat rare and frequent ICD …
[HTML][HTML] 视觉知识: 跨媒体智能进化的新支点
杨易， 庄越挺， 潘云鹤 - 2022 - cjig.cn
摘要回顾跨媒体智能的发展历程, 分析跨媒体智能的新趋势与现实瓶颈, 展望跨媒体智能的未来
前景. 跨媒体智能旨在融合多来源, 多模态数据, 并试图利用不同媒体数据间的关系进行高层次 …
前景. 跨媒体智能旨在融合多来源, 多模态数据, 并试图利用不同媒体数据间的关系进行高层次 …
Challenges of Real Life Few-Shot Image Classification
E Bennequin - 2023 - hal.science
In 2015, while deep neural networks achieved super-human performance in large-scale
image recognition, we started observing that this performance could not be reproduced with …
image recognition, we started observing that this performance could not be reproduced with …