Objectives
TEAMING.AI aims to overcome the lack of flexibility as a limiting factor of current Industry 4.0 while ensuring the role of the human being in the future industrial scenario by means of a human centred AI collaboration.
Knowledge Graphs
+
Relational Machine Learning
The project will rely on the combination of advanced methods for the representation of complex manufacturing processes by means of a novel approach which combines
knowledge graphs and
relational machine learning
to realise true human-AI teaming working schemes, thus answering the actual needs of the industry.
In order to realise this ambition TEAMING.AI will design and implement a novel human- AI teaming software platform composed of interacting utilities (with a knowledge graph at its core) supporting:
Auditable Ethics
The adoption of the ethical principle of human autonomy by design
An auditable model of trust in human-AI interactions based on committed roles and process models
Agile Development
Self-organizing and cross-functional teams of human agents and AI components
Novel software plattform for agile AI system engineering and operation
Operational Performance
Eased development and operation by enriched representations of processable knowledge
Advanced data analytics and optimization in dynamic manufacturing environments
Three Manufacturing Scenarios
1
2
3
TEAMING.AI’s paradigm will be particularised and materialised in three manufacturing scenarios with different challenges and requirements in terms of AI.
Agile production with high diversity of products and high frequency of process changes (quality inspection)
Knowledge-intensive processes such as process diagnostics for complex machines (injection moulding)
Harm prevention in challenging human-machine interaction scenarios (handling large-sized parts for machining)
Three Manufacturing Scenarios
1
2
3
TEAMING.AI’s paradigm will be particularised and materialised in three manufacturing scenarios with different challenges and requirements in terms of AI.
Agile production with high diversity of products and high frequency of process changes (quality inspection)
Knowledge-intensive processes such as process diagnostics for complex machines (injection moulding)
Harm prevention in challenging human-machine interaction scenarios (handling large-sized parts for machining)
AI Centric vs Human-Ai Teaming
Involved Humans
Operators
OP
Domain Exports
DO
Machine Learning Exports
MLE
Development Operators
DevOps
Contextualisation by KG (Knowledge Graph)
Agile Human in the Loop
