Artificial Intelligence Living Lab: Interactive Model Trainer workshop

One of Intelcomp's tools is the Interactive Model Trainer (IMT), which is the heart of the platform. The IMT is a tool specifically designed for training Artificial Intelligence models, enabling a user-in-the-loop approach that allows the incorporation of domain-specific expert knowledge into the models. Within IntelComp, its use is contemplated as part of the work of the living labs, for the selection of subsets of documents of interest in the scope of each living lab, as well as for the thematic characterisation of the selected documents.

Last June 6th, within the Artificial Intelligence living lab organised by SEDIA an online workshop focusing on the IMT took place. The event had a double objective: on the one hand, to present the main IMT functionalities to IntelComp stakeholders and, on the other hand, to generate a space for interaction where these members could play a key role in the final definition of IMT functionalities.

30 people attended the event, including stakeholders interested in learning more about the tool and IntelComp partners. The coordinator of AI Living Lab, David Gago (SEDIA), was in charge of presenting the event and explained the basic concepts of IMT, highlighting the added value of the elements of the tool. Subsequently, Joseba Sanmartín (FECYT, project coordinator) also talked about some general questions about the IntelComp project in general and IMT in particular. On the other hand, Asier López (NTT DATA) organised a co-creation dynamic, to facilitate interaction between the attendees, in which they contributed their different points of view.

The last part of the session was dedicated to generating a space for reflection and interaction about IMT under the prism of the Value Proposition Canvas methodology. IMT is already in a very notable degree of completion so the following events organized by SEDIA will focus on other tools, such as STI Viewer