A comprehensive survey of existing interpretation and explanation methods specifically designed for CapsNets.
Analyzes the feature representations learned by various capsule-based network architectures in different computer vision tasks.
Proposes a GRU-based sentiment analysis model for Arabic text using multilingual embeddings.
Investigates how Matrix Capsule Networks encode information and the interpretability of their internal representations.
Uses deep learning and ensemble methods to detect emotions in Arabic social media text.
Applies transfer learning with deep convolutional networks for automatic COVID-19 detection from X-ray images.
Presents a multilingual offensive language detection model for social media using ensemble and BERT fine-tuning.
Applies transfer learning and XGBoost to detect aggressive language in multilingual social media posts.
Introduces an Arabic dataset for commonsense sentence validation tasks.
Provides a benchmark dataset for Arabic commonsense explanation tasks.
Uses deep learning to classify human emotions from physiological signals.
Proposes a linked open data approach to resolve conflicts of interest in academic peer review.
Explores link prediction methods to recommend reviewers while mitigating conflicts of interest.
Summarizes recent methods for affect detection in Arabic textual datasets.
Compares LSTM-CRF models for Arabic named entity recognition tasks.