Catégories : Tous - performance - usability - ai - privacy

par Lucious Wang Il y a 24 jours

24

Software Architecture for Responsible Artificial Intelligence Systems: Practice in the Digitization of Industrial Drawings

The process of digitizing historical industrial drawings presents several challenges, such as the complexity of formats and the need for domain-specific expertise, which can lead to high error rates in manual digitization.

Software Architecture
for Responsible Artificial
Intelligence Systems:
Practice in the Digitization
of Industrial Drawings

Software Architecture for Responsible Artificial Intelligence Systems: Practice in the Digitization of Industrial Drawings

6. Conclusion and Future Work

Future improvements (interaction methods, expanded AI stack)
Expanding AI models to support various industrial applications
More interactive options (e.g., voice control)
Benefits of AI in drawing digitization
Reduces time and errors in industrial drawing digitization

4. Case Study: Piping & Instrumentation Diagram (P&ID) Drawing Digitization

Performance analysis: accuracy, time, availability
Application in digitizing historical piping & instrumentation diagrams

2. Key Components

Quality attributes (Usability, High performance, Accuracy)
Security & Privacy: Protection through federated learning
Performance: Efficient processing for large, multi-page drawings
Usability: Non-AI professionals should navigate easily
Role of AI technologies (OCR, computer vision, etc.)
Knowledge Graphs for relational data
OCR for text extraction
Computer Vision for element detection
Responsible AI software architecture
Responsible AI Principles: Aligns with EU's seven AI ethics principles (e.g., transparency, accountability).

5. Evaluation and Feedback

Federated model accuracy improvement
Limited samples for specific element training
Adapting to diverse drawing standards
End-user satisfaction (usability, interactivity, etc.)
Performance Gains: P&ID tasks completed in days instead of months
Interactivity, usability, and interpretability scored highly

3. Architecture Design

Collaborative multimodel fusion
Integrates specialized AI models (e.g., text and object detection)
Federated Learning for privacy
Federated Learning: Secures data by training models locally
Architectural Patterns: MVC, Layered Architecture
Layered Architecture: Separates concerns, enhancing usability and style adaptability
Model View Controller (MVC) for modular design

1. Introduction

Challenges in digitizing historical drawings
Complexity in historical drawing formats, domain-specific knowledge, and high error risk in manual digitization.
Purpose of responsible AI in industrial drawing digitalization