Between 2013 and 2016, our company had the opportunity to participate in the EU-funded FP7 R&D project, METALOGUE, as one of 10 Consortium Partners. The project aim was to develop a natural, flexible and interactive Multi-perspective and Multi-modal Dialogue system with meta-cognitive abilities, a system that can:
- monitor, reason about and provide feedback on its own behaviour, intentions and strategies and on the dialogue itself,
- guess the intentions of its interlocutor
- and accordingly plan (and if needed adapt) the next step in the dialogue.
The 3-year project ended in October 2016, had a budget of €3,749,000 (with the EU contributing €2,971,000) and brought together 10 Academic and Industry partners from 5 EU countries (Germany, Netherlands, Greece, Ireland and the UK).
METALOGUE focused on interactive and adaptive training situations, where debating and negotiation skills play a key role in the decision-making processes. Pilot systems were developed to train both types of skill.
Reusable and customisable software components and algorithms were developed and integrated into a prototype platform, which provides learners with an interactive environment that motivates them to develop meta-cognitive skills by stimulating creativity and responsibility in the decision-making, argumentation and negotiation process.
The project produced virtual trainers, Avatars, capable of engaging in natural interaction in English (with the possible addition of German and Greek in the future), using gestures, facial expressions and body language.
The pilots were evaluated by English-speaking students at the Hellenic Youth Parliament and several participating University campuses. Various industry verticals were also targeted in the course of the
project, in particular Contact Centres, e.g. to semi-automate and
enhance Call Centre Agent Training.
And here's the METALOGUE Avatar in action!
In this video, our full-body METALOGUE Avatar is playing the role of a business owner, who is negotiating a smoking ban with a local Government Counsellor
. Imperfect (e.g. some latency before replying and
an embarrassing repetition at some point!), but it also exhibits realistic facial expressions, gaze, gestures and body language and even selective and effective pauses
. It can process natural spontaneous speech in a pre-specified domain (smoking ban) and has achieved an ASR error rate below 24%
from almost 50% in the 1st year of the project!).