BERREKIN Project, in-circuit operators for natural language interaction in advanced manufacturing environments
Vicomtech coordinates the project, in which incremental learning techniques are being explored to facilitate the evolution of voice interfaces in industrial environments based on the expert knowledge of the operators.
Global, European and local industry strategies share a number of trends, such as the use of digital technologies and the promotion of collaboration between the two. The focus of the new production models is on the interaction between people and machines, in a context where the role of operators is key.
Aim of the project: towards natural language interaction systems with expert knowledge of factory operators.
The consortium formed by Vicomtech, Tekniker, UPV/EHU, Ikor Technology Center, IMH and UZEI, has been working for the last two years on facilitating the implementation of natural language interaction interfaces between operators and industrial production machines, based on the information contained in existing technical documentation. It has been found that, thanks to the latest technological advances, such interaction can be very feasible and advantageous, in that it is "hands-free" and/or "eyes-free" and requires minimal training, allowing operators to continue with physical tasks while communicating flexibly with the machine in question.
However, interaction systems derived from existing technical documentation have a major limitation: they are static and do not capture the expert knowledge of factory operators, which is not found in manuals. The main objective of the project is to address this limitation, through research and development in incremental learning technologies for conversational human-machine interactions, in scenarios where industrial operators can be in the loop, allowing conversational systems to evolve and adapt to the context of each industrial plant, ensuring a localised capture, storage and transfer of operators' expert knowledge.
Technologies that are being applied throughout the project
- R&D in incremental learning techniques for dialogue systems and in hybrid learning and exploitation policies that allow for greater flexibility in the systems.
- R&D in incremental learning techniques for question-answering (Q&A) systems.
- R&D on the application of the "Operator in the Loop" paradigm in distributed computing environments or Edge Computing for conversational human-machine interactions.
- R&D on interfaces to involve the "Operator in the Loop" to adapt, personalise and provide knowledge to conversational human-machine systems.
- Compilation of resources for Basque to implement "Operator in the Circuit" systems and conversational human-machine interactions also in that language.
Vicomtech's role: coordinators of the project
Vicomtech is currently coordinating the project, researching and developing incremental learning techniques for dialogue systems and question-answer systems, as well as exploring "operator-in-the-loop" distributed computing architectures.
In conclusion, thanks to the use of the most advanced Artificial Intelligence technologies and the close collaboration between the members of the consortium, BERREKIN project will be able to produce improvements in the optimization of productivity between operators and machines, which will undoubtedly have positive consequences in many industrial processes.
This project has been financed by SPRI within the framework of the Elkartek Program that supports collaborative research carried out by RVCTI agents in the RIS3 Euskadi strategic areas.