Aiming at the problems of low computational efficiency and insufficient accuracy in the privacy-preserving neural network based on homomorphic encryption， an efficient Homomorphic Neural Network （HNN） under three-party collaborative supporting privacy-preserving training was proposed. Firstly， in order to reduce the computational cost of ciphertext-ciphertext multiplication in homomorphic encryption， the idea of secret sharing was combined to design a secure fast multiplication protocol to convert the ciphertext-ciphertext multiplication into plaintext-ciphertext multiplication with low complexity. Then， in order to avoid multiple iterations of ciphertext polynomials generated during the construction of HNN and improve the nonlinear calculation accuracy， a secure nonlinear calculation method was studied， which executed the corresponding nonlinear operator for the confused plaintext message with random mask. Finally， the security， correctness and efficiency of the proposed protocols were analyzed theoretically， and the effectiveness and superiority of HNN were verified by experiments. Experimental results show that compared with the dual server scheme PPML （Privacy Protection Machine Learning）， HNN has the training efficiency improved by 18.9 times and the model accuracy improved by 1.4 percentage points.

Sepsis is a medical emergency triggered by pathogenic microorganisms such as bacteria， which can be life-threatening when severe， making early diagnosis and timely treatment crucial. In recent years， machine learning technology has shown tremendous potential in early prediction and treatment strategies for sepsis. By integrating data from multiple sources， machine learning models can assess patient risk accurately and identify high-risk individuals automatically， enabling early diagnosis of sepsis. In addition， machine learning can also assist physicians in developing personalized treatment plans. However， clinical applications based on machine learning methods still face a series of challenges， such as data standardization， model interpretability， and acceptance by medical personnel. Therefore， a comprehensive review was conducted on machine learning based sepsis prediction and intervention decision-making methods. Firstly， the basic process and framework of sepsis prediction and intervention decision-making were introduced. Then， the methods， relevant data and evaluation indicators of sepsis prediction and intervention decision-making were summed up systematically. Furthermore， a detailed summary of the specific applications of machine learning methods in sepsis-related clinical aspects was provided. Finally， the main challenges faced in this field currently were summarized， and future development trends were prospected.