All of the participants could relate to the topics of human scale intelligence and collaborative learning, and we agreed to use the workshop to consolidate ideas for joint research, and to outline a framework for joint research:
Understanding learning and knowledge transfer in a collaborative environment via cognitive computation
This overarching theme for the envisaged joint research was mapped out on the whiteboard.
Envisaged scope of the joint research
- To use the NeuCube technology to monitor at least 2 creative collaborators, who are geographically separated, (this could be a lab environment to start with) and connected to EEGs while wearing an immersive VR headset.
- We will record brain patterns as they communicate in a virtual world and sharing their respective mental models of a specific domain on a virtual canvas, and then reach formal agreement on a shared understanding of the domain using the MODA + MODE methodology and “validation via model instantiation”.
- We will analyse the 2 EEG recordings, run through the NeuCube to attempt to establish mappings between elements in the shared domain model and elements in the NeuCube representations derived from the two EEGs, to use these mappings for semantic and temporal analysis of the commonalities and differences in the two NeuCube representations of the domain. This may give us insights into the level of knowledge transfer between the two collaborators.
- This will provide us with a basis for further experiments in which we will explore both tacit and explicit knowledge creation and exchange, and give us a constructivist method for enhancing learning and creativity.
- As a second step we envisage developing a software tool that includes both NeuCube and S23M Cell Platform software for multilevel semantic modelling and knowledge representation to capture the knowledge shared as part of interactive learning.
- As a third step we envisage exploring learning and knowledge transfer between people with different cognitive lenses and different abilities for problems solving, and we may explore applications related to the teaching and learning of computer science.
Following the workshop, the scope of the research framework was formalised using the MODA + MODE human lens (see introductory slides for a simple example).
The envisaged research will make use of human scale semantic models for knowledge representation, AUT KEDRI’s NeuCube software, neural networks constructed with the help of NeuCube, S23M’s Cell Platform software, as well as new software tools that will be developed as part of the research.
The following list of relevant events is not complete, it is only intended to illustrate essential events in any process of interactive knowledge sharing and validation between humans.
The research will involve the following set of agents, including
- humans with various types of cognitive lenses,
- research team members from KEDRI and S23M,
- and neural networks created and simulated by NeuCube software.
Motivations and research objectives
The motivations of the joint research team relate to
- cognitive computation (symbolic),
- learning and knowledge transfer in a social environment,
- critical review of learning tools and techniques,
- and the design and development of software tools that enhance learning and creativity.
Formalising the shared understanding between agents
The following diagram illustrates how the level of shared understanding between two human agents can be formalised with the Cell Platform software, such that it becomes available for processing by software tools.
Over the coming months AUT KEDRI and S23M will jointly elaborate the above proposal.