7 September 2023 in Knowledge-Based Systems
9 November 2018 in International Journal of Fuzzy Systems
29 February 2024 in Knowledge-Based Systems
10 September 2025 in ACM Transactions on Information Systems
Cancer risk prediction is a cornerstone of personalized medicine that offers opportunities for early detection and preventive interventions. However, the current models are designed to predict cancer risk face several challenges. First, most rely on traditional statistical methods, which struggle to capture the complexity of genetic, family medical history, and lifestyle factors. Hence, the accuracy of these models is limited. Additionally, the models neglect to integrate multidimensional data sources, particularly genetic information like single nucleotide polymorphisms (SNPs), which could enhance prediction accuracy. Third, while the system might effectively predict risk, it cannot translate those predictions into actionable healthcare recommendations to reduce cancer risk. In this study, we address all three of these limitations. With a focus on six prevalent cancers—we extracted SNP data from the UK Biobank and designed a novel risk prediction model for cancer and personalized healthcare recommendations based upon the mixture of experts (MoE) paradigm and large language models (LLMs), respectively. Named MoE-HRS, experts based two router networks for separate processing by the Transformer and the convolutional neural network (CNN). Experiments on UK Biobank data show that our model outperforms state-of-the-art cancer risk prediction models. To bridge the gap between risk prediction and practical healthcare applications, we devised a healthcare recommender system powered by LLMs. This approach holds promise for enhancing early detection rates and promoting preventive healthcare management (relevant coding and data are available at https://github.com/bjtu-lucas-nlp/MoE-HRS ).
4 December 2020 in IEEE Access
Intelligent regulation for human exercise behaviors becomes significantly necessary for exercise medicine after the COVID-19 epidemic. The key issue of exercise regulation and its potential development for intelligent exercise is to describe human exercise physiological behaviors in a more accurate and sufficient manner. Here, a non-parametric modeling method with kernel-based regularization is presented to estimate cardiorespiratory biomarkers (i.e., oxygen uptake (V̇O2 ) and carbon dioxide output (V̇CO2 ) by merely non-invasively monitoring the indicator of exercise intensity (e.g., walking speed). Using the kernel-based non-parametric modeling, we show that V̇O2 and V̇CO2 behaviors in response to continuous and diversified exercise intensity stimulations can be quantitatively described. Furthermore, the dataset from the stairs experiment with a proper protocol is applied in the kernel parameter selection, and this selection approach is compared with the numerical simulation approach. The comparison results illustrate an improvement of 4.18% for oxygen uptake and 7.63% for carbon dioxide output in a half period, and 11.00% for oxygen uptake and 12.60% for carbon dioxide output in one period when using the kernel parameter selected from the stairs exercise. Moreover, the advantages of using the non-parametric model, the necessity of sufficient stimulation for identification and the importance of the kernel regularization term are also addressed in this paper. This method provides fundamental work for the practice of intelligent exercise.