Graph-Based Continual LearningBinh TangDavid S. MattesonInternational Conference on Learning Representations
Moreover, we propose a disentangle-based continual graph representation learning (DiCGRL) framework inspired by the human's ability to learn procedural knowledge. The experimental results show that DiCGRL could effectively alleviate the catastrophic forgetting problem and outperform state-of-the-art ...
2024 Cross-Regional Fraud Detection via Continual Learning With Knowledge Transfer TKDE 2024 Link Link 2024 Do not ignore heterogeneity and heterophily: Multi-network collaborative telecom fraud detection Expert Systems with Applications 2024 Link Link 2024 Connecting the Dots: Graph Neural Networks for ...
Adaptive scheduling for Internet of Vehicles using deconfounded graph transfer learning In this paper, we propose a Cross-City Federated Continual Learning Framework for Spatiotemporal Graph Transfer Learning called CCFTL, which removes ... X Liu,S Wang,Y Chen - 《Computer Networks》 被引量: 0发...
Urban development is a continual process of expansion and renewal (Li and Zeng, 2020). Urban renewal aims to improve built-up areas where current land functions and resource usage do not fulfill social and economic development requirements (Lee et al., 2016, YE, 2019, Du et al., 2021). ...
Explore related subjects Discover the latest articles and news from researchers in related subjects, suggested using machine learning. Knowledge Based Systems Genome assembly algorithms Robotics Social Robotics Robotic Engineering Surface Assembly ...
Continual learning for generative retrieval over dynamic corpora. In Proceedings of the 32nd ACM International Conference on Information and Knowledge Management, Birmingham, UK, 21–25 October 2023; pp. 306–315. [Google Scholar] Gao, Y.; Xiong, Y.; Gao, X.; Jia, K.; Pan, J.; Bi, ...
we propose an innovative robotic manipulation framework based on continual knowledge graph embedding. This framework enables hybrid robots to break free from the constraints of fixed rules set by human demonstrations, instead endowing them with inferring capability. The core idea is to utilize semantic...
A popular approach for predicting drug-target binding affinity is to feed the sequence of the target protein and drug (1D representation) into the deep learning model after it has undergone continual improvement. For example, DeepDTA [18] uses two convolutional neural network (CNN) blocks to lear...
learning capabilities. The authors present a K-means based 360-degree panoramic VR football training video distribution method. Furthermore, simulation studies are carried out using the content delivery network simulator, and the proxy server hit ratio, byte hit ratio, mean response time as well as...