DBIS

In our efforts to support heterogeneous communities with the tools and structures they need, we created our flagship peer-to-peer community platform las2peer. In our latest research efforts, we developed a social bot framework, which allows for a model-driven construction and utilization of social bots in the domain of technology enhanced learning (TEL). MobSOS, the community success measurement framework of las2peer, is used to aggregate distributed data from the microservices across the network and send it to the social bots, which use machine-learning technologies (e.g. RNN with TensorFlow) to compute their answer. As interfaces for interaction with the bot, we currently use both the learning application themselve (e.g. the bot reacts by acting as a participating user in the learning scenario), as well as a chat interface via Slack. Currently, the bot is only able to post their reaction messages to a channel, without the user being able to react to it. As an example application to evaluate our framework, we have used our informal learning application Distributed Noracle, as well as our las2peer service index.

The goal of this thesis is to utilize and evaluate the existing social bot framework with our existing, las2peer powered, community applications (e.g. the Requirements Bazaar). Additionally, the framework should be extended to support two-way interaction with the system via common chat interfaces (e.g. Slack).

If you are interested in this thesis, please do not hesitate to send a message to lange@dbis.rwth-aachen.de.