A5033745915- Emotion and Mood Recognition
- Advanced Graph Neural Networks
- Topic Modeling
- Face and Expression Recognition
- Text and Document Classification Technologies
- Data Quality and Management
- Video Surveillance and Tracking Methods
University of Electronic Science and Technology of China
2022-2024
<title>Abstract</title> With the innovation in computer vision, facial expression recognition (FER) is a dynamic research domain considering extensive practical applications various domains together with health, education, safety, law enforcement, Banking, marketing, and many more. The researchers have conducted tremendous work on basic expressions recognition, but less compound emotions which complex features due to combination of emotions. Different deep learning models been used for...
Few-Shot Knowledge Graph Completion (FKGC) aims to infer the missing facts of a relation with only few corresponding known instances. To accomplish this task, one prominent prerequisite is fully explore and exploit useful information from these supporting However, many existing methods usually treat support set as group independent contents, fail explicitly model interactions among They also majorly focus on positive instances in while overlooking negative ones. Generally, much has not been...
Unlike traditional knowledge graphs (KGs), which represent real-world facts as entity-relationship-entity triples, hyper-relational allow triples of entities with additional relation-entity pairs (also known qualifiers) to establish connections deliver more complicated messages. Modelling triplet-qualifier relationships effectively for accomplishing prediction tasks is an existing challenge. In this paper, improved graph completion method, STARE, proposed by introducing two new methods: (1)...