This workshop will take place at the Joint Ontology Workshop (JOWO) Episode V: The Styrian Autumn in Graz, Austria between 23-25 September 2019
This Second Workshop on INteraction-based Knowledge Sharing (WINKS-2) collocated with JOWO 2019 is aimed at researchers and practitioners investigating issues related to aspects of (autonomous) knowledge sharing, where the integration of knowledge is inherently interaction-based, irrespective of whether the interaction is machine to machine, or human to machine. Gradually expanding, distributed systems heighten the need of dynamic interactive knowledge-sharing processes and ever more sophisticated mechanisms are used to acquire and elicit knowledge. A paradigm shift has emerged that views knowledge creation, curation and evolution as a collaborative and interactive process between autonomous entities.
As a highly interdisciplinary workshop, WINKS-2 invites submissions that address the fundamental issues and challenges posed by interaction-based approaches to knowledge sharing. At the same time, we are interested in submissions that provide solutions for allowing knowledge sharing interactively, with a particular focus on the processes, mechanisms and protocols underlying the proposed solution.
An increasing number of intelligent systems communicate and exchange knowledge in various applications intended to facilitate our daily lives, from eCommerce and eBusiness systems that go beyond geographic and cultural borders, to intelligent personal devices that answer our every question at home or on the go. This multiplicity inevitably increases the diversity of such systems, which in turn poses new challenges to knowledge and data sharing within and across such systems. Heterogeneous computational entities (autonomous agents) generate and exploit knowledge collaboratively through some form of interaction. Interaction can occur between agents or with humans and typically involves dynamically exchanging information and reconciling different knowledge models as new or evolving knowledge sources are encountered. As ever more sophisticated (and interdisciplinary) mechanisms are used to acquire and elicit new knowledge, the effectiveness of traditional centralised sharing processes has become less tenable.
Present-day computer science, and Artificial Intelligence in particular, is facing a growing number of fundamental issues related to knowledge sharing, and approaches and techniques have been proposed to investigate and overcome them. A paradigm shift has emerged that views knowledge creation, curation and evolution as a collaborative and interactive process between autonomous entities.
Thus, interaction-based knowledge sharing has begun to receive significant attention as a way to address the challenges of interoperability between diverse systems in a timely manner. Its appeal is due to the fact that it has the same advantages and challenges as dynamic autonomous systems, yet its ambition emerges through the fundamental issues that it addresses. Interaction-based knowledge sharing allows distributed computational systems to seamlessly and intelligently combine their knowledge instead of imposing one knowledge model to all interacting parties, in an asymmetric relationship, therefore supporting the emergence of new knowledge. Furthermore, the feedback that occurs during interaction helps systems to control both the process and the success of the integration, and to better explain the outcome of the actions taken. However, this approach leads to several challenging requirements and assumptions, as the heterogeneity in vocabularies and methodologies requires the ability for agents to contextually adapt their policies, protocols and models. Additionally, the processing of emergent knowledge has to deal with novelty and unpredictable results. Finally, although humans are a valuable source of knowledge, the use of natural language as a knowledge and communication mechanism is still challenging and difficult to decipher.
By its very nature, interaction-based knowledge sharing is an interdisciplinary subject, grounded in various disciplines including: knowledge representation and reasoning; autonomous interaction and coordination models; qualitative negotiation and argumentation; trust and reputation; relation discovery; and knowledge graph acquisition, to mention only a few. This interdisciplinary problem appeals to, and transcends many of the paradigms the Autonomous Agents and Multi-agent community, but it extends to the wider Artificial Intelligence area and Information Systems in general. The first edition of WINKS involved researchers from multiple disciplines, and focused primarily on the ontology or vocabulary produced through interaction. In the second edition of WINKS we want to give more emphasis to the processing aspects, i.e., the theory and practice underlying interaction-based knowledge sharing. This makes the platform for discussion provided by the workshop especially appealing to the JOWO audience.