A divide and conquer framework for Knowledge Editing

Article Computer Science, Artificial Intelligence

GC-LSTM: graph convolution embedded LSTM for dynamic network link prediction

Jinyin Chen et al.

Summary: Dynamic network link prediction is a hot topic in network science, and our proposed model GC-LSTM, combining GCN and LSTM, can predict both added and removed links, making it more practical in reality.

APPLIED INTELLIGENCE (2022)

添加到收藏夹

Article Biochemistry & Molecular Biology

Toward better drug discovery with knowledge graph

Xiangxiang Zeng et al.

Summary: In this review, knowledge graph-based works for drug repurposing and adverse drug reaction prediction in drug discovery are summarized. The graph provides both structured and unstructured relations, while knowledge representation learning is a common approach to explore knowledge graphs.

CURRENT OPINION IN STRUCTURAL BIOLOGY (2022)

添加到收藏夹

Proceedings Paper Computer Science, Information Systems

Knowledge Graph Contrastive Learning for Recommendation

Yuhao Yang et al.

Summary: Knowledge Graphs (KGs) are useful in improving recommendation quality, but their sparse and noisy nature poses challenges. In this study, a Knowledge Graph Contrastive Learning framework (KGCL) is proposed to address these challenges by suppressing noise, deriving robust knowledge-aware representations, and utilizing supervision signals for contrastive learning. Experimental results demonstrate the consistent superiority of KGCL over state-of-the-art techniques in recommendation scenarios with sparse interactions, long-tail and noisy KG entities.

PROCEEDINGS OF THE 45TH INTERNATIONAL ACM SIGIR CONFERENCE ON RESEARCH AND DEVELOPMENT IN INFORMATION RETRIEVAL (SIGIR '22) (2022)

添加到收藏夹

Article Information Science & Library Science

A framework for creating knowledge graphs of scientific software metadata

Aidan Kelley et al.

Summary: An increasing number of researchers are using computational methods to generate or manipulate results in scientific publications. However, finding, setting up, and comparing scientific software is often challenging due to dispersed documentation and lack of structured metadata. This paper introduces a framework for automatically extracting metadata from scientific software documentation, structuring it into a Knowledge Graph (KG), and providing a framework for browsing and comparing the KG contents.

QUANTITATIVE SCIENCE STUDIES (2022)

添加到收藏夹

Review Biochemical Research Methods

Enriching contextualized language model from knowledge graph for biomedical information extraction

Hao Fei et al.

Summary: Biomedical information extraction (BioIE) aims to analyze biomedical texts and extract structured information, incorporating external structural knowledge can significantly improve performance, especially when integrating a large scale of biomedical knowledge graphs (BioKGLM).

BRIEFINGS IN BIOINFORMATICS (2021)

添加到收藏夹

Article Computer Science, Software Engineering

The Probabilistic Description Logic BALC

Leonard Botha et al.

Summary: Description logics are formalisms for representing terminological knowledge, but are not suitable for dealing with uncertainty. This paper introduces a Bayesian extension BALC of the propositionally closed DL ALC, presenting a tableau-based procedure for solving consistency and other inferences, showing that these problems remain ExpTime-complete like in the classical ALC.

THEORY AND PRACTICE OF LOGIC PROGRAMMING (2021)

添加到收藏夹

Article Computer Science, Interdisciplinary Applications

Drug repurposing for COVID-19 via knowledge graph completion

Rui Zhang et al.

Summary: This study utilized literature-derived knowledge and knowledge graph completion methods to identify potential drug candidates for COVID-19. The accuracy classifier based on PubMedBERT performed the best in identifying accurate semantic predications, and the TransE model outperformed others in predicting drug repurposing candidates. Several known drugs linked to COVID-19 were identified, as well as novel drugs that have not been studied yet.

JOURNAL OF BIOMEDICAL INFORMATICS (2021)

添加到收藏夹

Article Computer Science, Information Systems

Multilayer Internet-of-Things Middleware Based on Knowledge Graph

Cheng Xie et al.

Summary: This article proposes a knowledge graph-based multilayer IoT middleware that introduces a new layer to bridge the gap between devices with different communication protocols and can uniformly manage all devices. Evaluation of the proposed approach in a real-world project shows that it effectively resolves communication gap and heterogeneous access issues in the system.

IEEE INTERNET OF THINGS JOURNAL (2021)

添加到收藏夹

Article Computer Science, Artificial Intelligence

Relational Learning Analysis of Social Politics using Knowledge Graph Embedding

Bilal Abu-Salih et al.

Summary: Knowledge Graphs (KGs) have gained significant attention recently and have been applied to tackle various real-life problems in different domains. Despite the abundance of generic KGs, constructing domain-specific KGs is essential. This paper introduces a credibility domain-based KG Embedding framework and demonstrates its effectiveness through experiments.

DATA MINING AND KNOWLEDGE DISCOVERY (2021)

添加到收藏夹

Article Computer Science, Artificial Intelligence

DTN: Deep triple network for topic specific fake news detection

Jinshuo Liu et al.

Summary: Detection of fake news has attracted wide attention in various fields, and incorporating knowledge graphs is crucial for improving language understanding and providing explanations for fake news detection. The proposed deep triple network (DTN) leverages knowledge graphs to enhance fake news detection performance.

JOURNAL OF WEB SEMANTICS (2021)

添加到收藏夹

Article Computer Science, Information Systems

A Knowledge Graph Based Approach to Social Science Surveys

Jeff Z. Pan et al.

Summary: The recent success of knowledge graphs has sparked interest in applying them in open science, such as in intelligent survey systems for scientists, however, evaluating the quality of candidate survey questions provided by these methods has been limited. Proposed a dynamic and informative solution for an intelligent survey system based on knowledge graphs, allowing for selection of subsequent questions based on responses to previous questions.

DATA INTELLIGENCE (2021)

添加到收藏夹

Proceedings Paper Computer Science, Artificial Intelligence

Dynamic Routing Networks

Shaofeng Cai et al.

Summary: This paper introduces Dynamic Routing Networks (DRNets) to support efficient instance-aware inference, reducing parameter size and FLOPs while maintaining prediction performance comparable to state-of-the-art architectures.

2021 IEEE WINTER CONFERENCE ON APPLICATIONS OF COMPUTER VISION WACV 2021 (2021)

添加到收藏夹

Review Computer Science, Information Systems

Knowledge Graphs in Manufacturing and Production: A Systematic Literature Review

Georg Buchgeher et al.

Summary: Knowledge graphs in manufacturing and production can enhance production efficiency and quality output, aiding companies in achieving Industry 4.0 goals. However, further research is needed as existing studies in the field are still in early stages, with gaps in understanding how knowledge graphs can be effectively applied in manufacturing and production.

IEEE ACCESS (2021)

添加到收藏夹

Article Computer Science, Software Engineering

The realist approach for evaluation of computational intelligence in software engineering

Raghavendra Rao Althar et al.

Summary: Secured software development requires a security mindset throughout software engineering practices, with an emphasis on considering software security during the requirements phase. Utilizing concepts like machine learning to understand data sources can help improve objectivity in conversations between requirements gathering teams and customer business teams. Strengthening traditional methods like threat modeling and exploring feature engineering in software development can also enhance security focus in requirements gathering practices.

INNOVATIONS IN SYSTEMS AND SOFTWARE ENGINEERING (2021)

添加到收藏夹

Article Biochemical Research Methods