In the general research paradigm of BioNLP, corpora provide specific customized settings for entities, relationships, and topics in the metadata of knowledge entries. Corpus-based customized literature mining has rapidly developed over two decades, generating a large number of knowledge entries across multiple fields such as agriculture, biology, and medical health. Given the vast database of knowledge entries, ensuring the robustness of the results is crucial. In this talk, I will introduce several attempts by my research group in recent years to integrate multi-omics data with text knowledge entries for knowledge fusion. Here, data results from GWAS, RNA-Seq, etc., provide gene-phenotype associations and are used to enhance the associative evidence in text knowledge entries. This talk will cover how deep learning methods are used for semantic perception of knowledge entries and how Bayesian networks are applied for associative judgment and consistent fusion of knowledge entries from different sources. Additionally, I am happy to discuss some recent attempts and new ideas in knowledge integration aimed at enhancing the robustness of knowledge entries.

Bio:
Dr. Jingbo Xia is an associate professor in the Department of Artificial Intelligence, College of Informatics, Huazhong Agricultural University, China. He is also a member of the Bioinformatics Key Laboratory of Hubei Province, a member of the Engineering Research Center of Intelligent Technology for Agriculture, China, and a member of SIGBIOMED (ACL Special Interest Group on Biomedical Natural Language Processing). He currently acts as an associate editor in <Computational Intelligence>, and editorial board member in <Scientific Data>, <BMC Medical Informatics and Decision Making>, and <Journal of Mathematics Research>. His main research interests include 1) Corpus design and Biomedical knowledge discovery based on BioNLP; 2) Data mining for geno-phenotype association.