Computational systems pegged for a communication skills upgrade
The ESSENCE project was launched with the vision to establish a sustainable research base for training, joint research and commercial exploitation of new techniques in the area of shared, evolving semantics. The Marie-Curie Initial Training Network (ITN) conducted innovative research into the evolution and negotiation of meaning among human and artificial agents, noted project coordinator Dr. Michael Rovatsos. For four years to 2017, 15 early-career researchers studied semantic technologies, language games, multiagent communication, ontology learning and human dialogue to extend the research vision of diversity-aware artificial intelligence (AI). ‘This vision emphasises creating next-generation AI technologies that can be used to bridge the gap between heterogeneous agents by exploring how representation, reasoning and interaction can be used to allow diverse collectives of agents to share information and knowledge, coordinate their activities and combine their individual capabilities,’ explained Dr. Rovatsos. Establishing a research and training community ESSENCE assumed an innovative approach to interdisciplinary doctoral and postdoctoral training in a first-of-its-kind training programme for this emerging area. A broad range of training capabilities and resources across countries and different sectors and disciplines supported early-career researchers in their individual research projects. ‘The ITN hosted a series of events, provided open data sets and software that allow other researchers to engage with our work, and developed a challenge competition to allow broader communities to engage with the challenges we are interested in,’ said Dr. Rovatsos. The overarching aim of the 15 individual projects was to be able to replicate the capabilities that humans exhibit when negotiating meaning in AI systems. To this end, they applied methods from diverse research areas such as theoretical linguistics, multiagent systems, argumentation, semantic technologies, dialogue systems and language evolution. ‘Through extensive collaboration between different groups working in these areas, we identified diversity-awareness as the unifying characteristic of the methods we seek to bring together,’ added Dr. Rovatsos. Extending diversity awareness and communication skills Important progress was made on various problems related to the broader vision of diversity-aware AI. A key achievement, according to Dr. Rovatsos, was ‘the insight that a holistic conceptual framework for diversity-awareness has the potential to bring about a step change in AI’. This can be realised ‘by enabling the re-use and re-combination of single-purpose systems in complex ecosystems of heterogeneous, collaborating AI systems and people.’ To help build a sustainable European research community, ESSENCE carried out various activities, including ‘integration’ projects. The ESSENCE Platform project produced a publically available software repository to enable reuse of the code produced in the network. In the ESSENCE Model project, a collection of data sets and benchmarks were developed that allow researchers to use the software components or experiment with the data sets using their own methods. Finally, the ESSENCE Challenge was run as an open competition to engage people around the world who are interested in solving hard ESSENCE-related problems. Other activities included six workshops and three summer schools, and the project’s flagship community-building event, the Final ESSENCE Conference held in October 2017. Additionally, the ITN authored 70 scientific publications. Overall, the various activities ‘succeeded in putting diversity-awareness “on the map” in the European AI landscape, and fostered the emergence of a new community in the area,’ Dr. Rovatsos concluded. Further, ESSENCE designed a curriculum, training materials and a joint PhD programme structure that can be applied for future training in this research area.
Keywords
ESSENCE, communication, diversity-awareness, computational systems, AI systems, artificial awareness, semantics
