Industrial Artificial Intelligence - From Automated Process to Cognitive Analytics

Explore the transformative impact of industrial artificial intelligence on manufacturing, transportation, and supply chains in this 41-minute talk. Learn how AI technologies are leveraging vast amounts of data to optimize processes, predict failures, and revolutionize traditional operations and maintenance. Discover the potential for unprecedented savings, improved engineering efficiency, and enhanced safety through AI applications. Examine the challenges of implementing AI in industrial settings, including concerns about robustness and resilience. Delve into various decision-making models, from fully centralized to decentralized, and understand the rise of edge computing in industrial contexts. Investigate the importance of taxonomies, ontologies, and context awareness in different industrial scenarios. Compare data-driven and model-based approaches, and explore the process of hybrid prediction. Analyze the challenges in descriptive, diagnostic, predictive, and prescriptive analytics, including the concept of the "digital butterfly effect." Gain insights into the types of data analytics and the need for complete datasets to ensure effective AI implementation in industrial environments.

Intro
Scale up and populate.. The Achilles heel
Elements in the Ai process
The infrastructure
Fully centralized decision
Semi centralized decision
Fully Decentralized decision
What characterizes industry and transport systems?
Edge computing is on the rise in many industries
Traditional way, we transfer everything to cloud
With Multi agent for large fleet
Analytics based on OT
Taxonomies and ontologies
TRANSFORMATIVE MAINTENANCE SOLUTIONS Integration & Application of Technologies
The challenge in Descriptive analytics
What is context awareness?
Railway context
Marine context
Diagnostic analytics
The challenge in Diagnostics analytics
Black Swan Losses
Data driven or model based?
The process of hybrid prediction
Predictive analytics:RUL prediction and simulation of scenarios
Prescriptive challenge: The digital butterfly effect
Is my ML model enough?
Need for complete datasets
Types of data analytics