Multimodal Learning Analytics for Game-based Learning

23 Sep 2020 - Arun Balajiee

Speaker: Andrew Emerson

date: 09/23/2020

This week’s talk was very close and within the field of my research - education technologies. I had read a little about the Crystal Island, which is a game based environment about infectious diseases. Emerson’s work is an adaptation on top of the gap to support and understand the effects of multi-modal feedback in the interactive game.

Specifically the research questions of the Emerson et al. is to classify the level of interest and performance of students learning through playing the game. The aspect of learning analytics was in developing student adaptive content and game difficulty based on the strategies used by the students in their gameplay and capturing their interaction with facial features through eye gaze and eye tracking. Further, the gameplay is analyzed in a posttest session by the designer to get interactive trace logs.

Most of the first part of the talk was a discussion based on this submission this year, namely Multimodal learning analytics for game‐based learning. In the paper they discuss the technical details and the results of the implementation in much detail. Most of my take away from the talk was in terms how he developed the ideas to approach the problem of understanding student affect in learning and the approach to solving the problem thourgh multimodal learning data analytics. Indeed, why specifically did the author consider try understanding the data using multimodal analytics.

The later parts of the talk got more interesting when this idea was extending in the implmentation of multimodal analytics in musuems, that is in informal learning ecosystems. There are several challenges in this setup in developing a study design, the simulation of different scenarios to solve the problem and such. The author discussed firstly how they solved hte problem of developing a model to process and classify the interaction data into different levels of interest. Further, the author discussed the developement of study design for information on interaction data and the learning gains from these interactions. The techincal details are discussed in further detail in their AIED publication

I found both parts of the talk really engaging. I found really interesting branches that I could utilize as scaffolds or prompts to generate ideas for my own research in the field of educational technologies. This is certainly one of the best talks that I have attended recently in which I could understand and relate to some of the results and put myself from the viewpoint of the speaker to be able to extend out of the ideas that I could gather from the discussion during the talk.