Meeting Minutes
===
###### tags: `AI Proficient` `Meeting' 'INEOS' 'WP1'
Nov 10th, 2020, online

* Alexandre Voisin (UL)
* Benoit Iung (UL)
* Marc Anderson (UL, Post-Doc Ethics)
* AAlaeddine Chaoub (UL)
* Paul Astiazaran (Conti)
* Aitor Arnaiz (Tekniker)
* Santiago Fernandez (Tekniker)
* Uros Milosevic (TenForce)
* Nikola (IMP)
* Lazar Berbakov (IMP)
* Dea Pujic (IMP)
* Christophe (INEOS Coordinator)
* Patrick Primus (Ineos Geel - process control - polyprophilene)
* Bram Auwers(Ineos Geel - process control)
* Suzanne Knuttnick-Conrad (Ineos Cologne - quality manager - polymer)
* Marc Ruby (Ineos Cologne)

Christophe Introduction
What can Ineos expect from the project
in the short and long term?
and workload?

Two 14 member groups
What does each of the partners add to the project?

# Roundtable
see video as well
Christophe - In principle not use cases on PdM at Ineos
No trouble.
Cologne - Process control...
Not all the "solutions expected" are necessarily to be experimented on the 3 sites. Each site can support specific considerations with regards to the type of product (e.g continuous, discrete), types of processes, abstraction levels (ex. edge, platform) ...
On the other hand, it is important that all the "solutions" proposed and developed are at least experimented on one site. Of course, if a solution can be customised to each site, this will make it possible to conclude on the degree of generalisation offered by the solution, which is a criterion expected by the commission [Keep in mind the general concept defended by the project. See section 1.3.1.]

# WP 1

Human at the center of the call
Have the general concept of AI-PROFICIENT, define use cases for the plants. Maybe some are common?
The aim is to demonstrate the solution at least on one demonstrattion plant.
Customize solution to

discrete (Continental )vs continuous processes (INEOS)

Would be good if product could be used in different domains

## T1.2

Human can feel in competition with AI
Ethics as value
bottom up approach
Special Session on Ethics helpful (Ineos)

## T1.3

Need to understand the context of the use cases th
Need also to understnad if the technology solutions can be applied to the use case challenges
-lots of data available

## T1.4
- define KPIs
- define project requirements

## T1.5
system architecture
- a preliminary sketch of the architecture - to be followed as much as possible
- specify interfaces and services to be developed
- inputs and outputs?
-------------------------
define the use case over the first year (first 6 months) ===>> yes - agree - first 6 months! IMPORTANT
Have a view of the solution provided (end of first year)

## Continental presentation + Use case template
- continental always wants human in Command
- need to define KPIs which will hold for the next three years
- is the Use Case ethical enough?
- Define set of tags in relation to the project issues like: human related, predictive maintenance, Digital twin, process optimization... (??)
- classification of the Use Cases is important : predictive maintenance or human interaction classification of the Use Cases is important : predicitve maintenance or human interaction
Paul: Could be interesting to characterise and standarise and classify well the different cases
- add information related to the project and overall concept (later):
- edge/cloud level,
- to which STO's is the UC related to?

# Presentation of INEOS
delayed to 16/11

Discussion about project organization and meeting scheduling

extra meetings as necessary between smaller groups in the project