Search

Transfer learning (TL) is a machine learning (ML) method in which knowledge is transferred from the existing models of related problems to the model for solving the problem at hand. Relational TL enables the ML models to transfer the relationship networks from one domain to another. However, it has two critical issues. One is determining the proper way of extracting and expressing relationships among data features in the source domain such that the relationships can be transferred to the target domain. The other is how to do the transfer procedure. Knowledge graphs (KGs) are knowledge bases that use data and logic to graph-structured information; they are helpful tools for dealing with the first issue. The proposed relational feature transfer learning algorithm (RF-TL) embodies an extended structural equation modelling (SEM) as a method for constructing KGs. Additionally, in fields such as medicine, economics, and law related to people’s lives and property safety and security, the knowledge of domain experts is a gold standard. This paper introduces the causal analysis and counterfactual inference in the TL domain that directs the transfer procedure. Different from traditional feature-based TL algorithms like transfer component analysis (TCA) and CORelation Alignment (CORAL), RF-TL not only considers relations between feature items but also utilizes causality knowledge, enabling it to perform well in practical cases. The algorithm was tested on two different healthcare-related datasets — sleep apnea questionnaire study data and COVID-19 case data on ICU admission — and compared its performance with TCA and CORAL. The experimental results show that RF-TL can generate better transferred models that give more accurate predictions with fewer input features.

COVID-19 has spread to all countries around the world after it was first discovered in Wuhan of China at the end of 2019. It is caused by a novel coronavirus called SARS-CoV-2 with much semblance to SARS-CoV including the sequence homology and disease symptoms. It is reported to be more infectious than SARS-CoV due to higher binding affinities between its spike protein and the ACE2 receptor on cell surface. Despite this, its case fatality rate is much lower compared with that of SARS-CoV although it varies in different countries. However, the case fatality rate increases steadily with age and it is reported to be the highest in aged COVID-19 male patients in almost all countries. Consistent with these, females have higher antiviral immune responses. Males and females are different in inflammatory response and aberrantly hyperactive cytokines are the main lethal causes of COVID-19. Interestingly, the gene encoding the ACE2 receptor protein and some genes encoding the immune regulatory proteins such as TLR7 are located on X chromosome which is subject to X chromosome inactivation and sex hormone regulation. These may account for some sex-dependent immune responses and lethality observed in COVID-19 patients. In general, children are less likely to be infected with SARS-CoV-2 and only less than 1% of pediatric COVID-19 patients may die of COVID-19. However, the most severe pediatric cases become multisystem inflammatory syndrome that is similar to Kawasiki disease with features of viral infection. Since most infected kids were boys in China, there may be sex-dependent immune response in pediatric COVID-19 cases as well.