About Forum Brain



Forum Brain is a technology developed and tested over several years by Associate Professor Simon Angus from SoDa Laboratories and the Department of Economics at the Monash Business School, Monash University, Melbourne, Australia.

Whilst open-ended online discussion forums are highly valuable for online learning delivery, analysing the quality of student contributions to a forum is a notoriously complex and labour intensive task for educators. Leveraging state of the art technologies from natural language processing, graph theory and machine learning, Forum Brain provides educators with a scientifically valid, and entirely automated tool that provides key measures of student activity, engagement and quality based on their contributions to forums.

Forum Brain’s quality of contribution analysis has been trained and validated on experienced human grader labelling of hundreds of real threads. Using ensemble methods, Forum Brain achieves high accuracy in determining the best contributors to any thread, with new technologies constantly being harnessed to improve its predictions.

Currently, Forum Brain supports Moodle (v.3.5). Users can analyse any forum by backing up the forum within Moodle, and then uploading the mbz backup file to their profile area on Forum Brain. Forum Brain will automatically analyse the uploaded forum file and provide the user with a student-level set of analytics as a comma-separated-value (csv) file for either direct download or email to the user’s account. The report is also saved to the user’s profile for future use.

For use notes about the output from Forum Brain and to view an example report, please get in contact.

Forum Brain was developed to support fair, insightful, and rich assessment of student engagement through online forums.

Forum Brain has been trained specifically to answer the following question: “Who made the most valuable contribution to this thread?”. The question is intentionally open-ended in order that regardless of topic or domain, great contributions to the discussion are recognised and rewarded.

Grading contributions to forums is a difficult task for educators, and unfortunately, educators are not always able to provide the attention required to each and every post you might make, in its context. Forum Brain, being an intelligent, automated method, treats every thread in exactly the same way.

Key design features include :

  • Intelligent: Forum Brain uses a very large number of features from your contributions in context, and others which are mapped from more abstract representations of your contributions, meaning that a rich information set is always taken into account.
  • Consistent: Forum Brain doesn’t get tired or rushed. Its analytic engine approaches every single contribution in exactly the same way.
  • Unbiased: Forum Brain is specifically designed to be gender and ethnicity blind.
  • Reasoned: Forum Brain’s analysis is based on transparent weighting of the key attributes of a given set of thread contributions to their value in the wider discussion.
  • Game-proof: Forum Brain’s complex and broad information set is impossible to game. Good students will be rewarded and lazy contributors will not.
  • Fair: Taken together, Forum Brain puts premium on fairness, something all students should value.

Associate Professor Angus Leads the Forum Brain initiative and is an active researcher and technologist in complexity science, data science, and computational economics, and is also one of the most highly awarded educators in the Australian tertiary system, having being recognised multiple times by his peers with a national tertiary citation and award for excellence for his innovative, engaging and passionate approach to learning and teaching. Associate Professor Angus is a also co-founder of SoDa Laboratories, an alternative social data science group at Monash University; KASPR Datahaus, an alternative data company; and the IP-Observatory, providing global internet quality and activity measurements.

Mr Satya Borgohain is the lead web developer for Forum Brain and is currently a Research Fellow at SoDa Laboratories. He has a keen interest in machine/deep learning paradigms and is conducting research on biologically plausible neural networks along with applied NLP in Social and Political science domains. His prior experiences include two years as a full stack engineer where he implemented and deployed predictive models and as a computer vision engineer at a tech startup which provides aid to the elderly in aged care. He is deeply passionate about programming and the broader impacts of technology in general and more specifically, engineering intelligence.

Back to Home