撰写流程

1. 首页:
   • 页眉: 公司徽标或logo
   • 标题
   • Version: 1.0 | Jan 2020
   • Prepared by: Person 1 | Reviewed by: Person 2
   • Revision History

     Version	Author(s)	Description of Version	Date
     1.0	Person 1	Initial document drafting	13 Feb 2019
     1.1	Person 2	Chapter 3 (3.1, 3.2, 3.3)	16 Feb 2019

2. 第二页
   • 目录: ​Table of Contents
3. 之后
   • 正文

目录模板

模板1

1. Introduction
    1. Project Purpose
    2. Project Deliverables
    3. Technical Challenges and Proposed Solutions
    3.  
    4. State-of-the-art Technigues Integration   
    5. Novel Features 

2. Technical Preliminaries 
    1. Facial Recognition
    2. Adversarial Attacks and Defenses 
    3. Face Swapping and Detection 

3. Platform System Architecture
    1. Facial Recognition System
        1. FaceNet Model
        2. VGG-Face Model
    2. Adversarial Attack System
    3. Adversary Detection System
    4. Face Swapping System
    5. Face Swapping Detection System

4. Software Architecure Design 
    1. Overall Architecture 
    2. Adversary Attack Tesing Module 
    3. Adversary Attacks Detection and Forensics Module
    4. Face Swapping Detection and Forensics Module 
    5. Intermediate Transmission Modules
    6. GUI Analysis and Visualization module
    7. The other module 

5. Conclusion 
    1. Project Achievements
    2. Future Research 

- References

- Appendix I 

3. ASTRI Adversarial Face Detection System 
    1. System Architecture 
    2. Facenet Recognition System 
        1. Face Detection
        2. Face Alignment
        3. Face Comparison
        4. Face Recognition
    3. Adversarial Attack System 
        1. Threat Model
        2. Adversarial Attack Theory
        3. Implement Adversarial Attack
        4. Adversarial Attack Result
    4. Adversarial Face Detection System
        1. Face 68 landmarks Location
        2. VGG-Face Model Architecture
        3. Attribute Neurons Extraction
        4. Augmented Model Implementation
        5. Adversarial Face Detection

4. Face swapping and detection system
    1. System Architecture 
    2. Face Swapping System 
        1. FFmpeg 
        2. Face Location 
        3. Face swapping 
    3. XceptionNet Detecting System
        1. Xception Model Architecture 
        2. Automatic Xception Detection Method 
    4. FWA(Face Warping Artifacts) Detection System
        1. Face Warping Artifacts Method 
        2. Method 

- Testing and Forensics Results  
    1. Test cases
        1. Case I: Adversarial attack testing 
        2. Case 2: Adversarial face detection testing 
    2. Results Analysis

模板2

1. Introduction

2. System Architecture
    1. System Architecture Design
    2. Web Frontend Portal Server
    3. Web Admin Portal Server
    4. OCR Engine Server
    5. Data Management and Analytics Server
    6. Credit Assessment Engine Server
    7. Loan Management Server
    8. User Management Server
    9. Identity and Access Management Server
    10. Secure Data Server
    11. API Partner Servers
    12. Database Servers

- APPENDIX I
    1.	System Process Flow
    2.	Application Process Flow
    3.	API Specifications
    4.	UML Sequence Diagrams
    5.	Define Rule for OCR Engine

- Data Analytics Engine for Credit Assessment and Monitoring
    1. Framework of Data Analytics Engine 
    2. Datasets of Data Analytics Engine
    3. List of Machine Learning Algorithms
    4. Model Output Visualization

- White Paper for Proposing a New Alternative Credit Assessment Framework

模板3

- Acknowledgements
    | No. | Member of the Advisory Panel |       Representatives       |
    |-----|------------------------------|-----------------------------|
    |  1  |    Bank of China Hong Kong   |Dr xx, Ms xx, Ms xx and Mr xx|
- Foreword
- Executive Summary

- Part One: The creditworthiness of micro-, small and medium-sized enterprises: a 360-degree view
    1. Challenges and opportunities relating to credit assessment for MSMEs
        1. Micro-, small, and medium-sized enterprises in Hong Kong
        2. The conventional credit assessment approach
        3. Key challenges in assessing MSME credit risk
        4. Adoption of artificial intelligence and machine learning for alternative credit assessment
    2. Data for alternative credit assessment
        1. Classification of data for evaluation of creditworthiness
        2. Transactional data
        3. Non-transactional data
        - Featured section: An example of a psychometry assessment from CRIF
        4. Benefits and challenges for lenders of using alternative data for credit assessment
    3. Industry examples of alternative credit assessment
        1. The global landscape of MSME loan lenders in practicing alternative credit assessment
        2. Case studies in the implementation of alternative credit assessment

- Part Two: Fintech infrastructural components for alternative credit assessment
    1. Development of machine learning models for default prediction
        1. Model Selection
        2. Data Exploration
        3. Model Training
        4. Model Assessment
        5. Feature Importance
        6. Model Interpretability
    2. Automation of credit underwriting for alternative credit assessment
        1. Online lending platform for the automation of credit underwriting
        2. Workflow for Alternative Credit Assessment
        3. Combining conventional and alternative credit assessment models
        - Featured section: High-level machine learning-based credit scoring framework for MSME lending
    3. Privacy-enhancing technologies in sharing alternative credit data
        1. Differential privacy
        2. Zero-knowledge proof
        3. Secure multi-party computing and homomorphic encryption
        4. Federated Learning
        5. Evaluating MSME credit ratings with federated analysis, homomorphic encryption, and differential privacy
    
- Part Three: Deploying alternative credit assessment
    1. Evaluating the performance of machine learning algorithms on MSME data
        1. Setup of the experiments
        2. Results of the comparison of machine learning algorithms
        3. Insights gained from applying machine learning algorithms to MSMEs’ financial data
        4. Case Study: Credit assessment of Japanese MSMEs
    2. An industry-specific alternative credit assessment framework
        1. Proposed framework for alternative credit assessment
        2. The demonstration of the Retail Alternative Credit Assessment (RACA)
    3. Roadmap ahead
        1. Facilitation of data availability
        2. Continuous technical advances in modelling
        3. Centralised data-sharing platform for alternative credit assessment

- Appendix A - Operational considerations in the deployment of the alternative credit assessment
    1. Section 1: Background	
    2. Section 2: Regulatory construct	
    3. Section 3: Government leadership	
    4. Section 4: Machine learning and AI	
    5. Section 5: Data subject consent	
    6. Section 6: Building an ecology	

- Appendix B - Machine learning algorithms for model training and default prediction
    1. Section 1: Ensemble learning techniques
    2. Section 2: Common machine learning algorithms
        1. Logistic Regression
        2. K-Nearest Neighbours (KNN)
        3. Random Forest
        4. Extra-Trees
        5. CatBoost
        6. LightGBM
        7. XGBoost
        8. Convolutional neural network (CNN)
        9. Stacking

模板4

1. Introduction

2. Architecture
    1. System Architecture
    2. Android mobile app
    3. Dispatcher Server
    4. Token server
    5. Verification server
    6. Risk management server
    7. User server
    8. Database table

3. Framworks for Biometric Authentication
    1. Facial recognition
        1. Face Detection
        2. Face feature extraction and landmark detection
        3. Threshold configuration
    2. Speaker recognition
    3. Handwritten signature recognition

4. Conclusion

- Appendix I
    - System Process Flow
        - System Process Flow for Onboarding
        - System Process Flow for Login
        - System Process Flow for Biometric Verification
        - System Process Flow for Banking Service 
    - APP Flow Diagram
        - Onboarding Cross-Functional Process Flow
        - Login Cross-Functional Process Flow
    - API Specification
        - API Overview
        - Function Logic – User Onboard
        - API Sequence Diagram – User Onboarding
        - Function Logic – User Login
        - API Sequence Diagram – Login
        - Function Logic – Banking Service
        - API Sequence Diagram - Banking Service