Welcome to TrueLearn’s documentation!

TrueLearn is a machine learning library for modelling lifelong learner engagement with educational resources.

Get Started

See the Getting Started for a quick introduction to the package.

Related Papers

This work is related to the following papers:

• TrueLearn A Family of Bayesian Algorithms to Match Lifelong Learners to Open Educational Resources published at the Thirty-forth AAAI Conference on Artificial Intelligence, 2020 at New York, NY, USA. BibTex below:

  @inproceedings{truelearn2020,
      author = {Bulathwela, S. and Perez-Ortiz, M. and Yilmaz, E. and Shawe-Taylor, J.},
      title={TrueLearn: A Family of Bayesian Algorithms to Match Lifelong Learners to Open Educational Resources},
      booktitle = {AAAI Conference on Artificial Intelligence},
      year = {2020}
  }
  

• Power to the Learner: Towards Human-Intuitive and Integrative Recommendations with Open Educational Resources published in 2022 as part of the Special Issue on AI and Interaction Technologies for Social Sustainability. BibTex below:

  @Article{su141811682,
     author = {Bulathwela, Sahan and Pérez-Ortiz, María and Yilmaz, Emine and Shawe-Taylor, John},
     title = {Power to the Learner: Towards Human-Intuitive and Integrative Recommendations with Open Educational Resources},
     journal = {Sustainability},
     volume = {14},
     year = {2022},
     number = {18},
     article-number = {11682},
     url = {https://www.mdpi.com/2071-1050/14/18/11682},
     ISSN = {2071-1050},
     DOI = {10.3390/su141811682}
  }
  

• PEEK: A Large Dataset of Learner Engagement with Educational Videos published in 2021 as part of the 4th workshop on Online Recommender Systems and User Modeling at ACM RecSys 2021. BibTex below:

  @article{DBLP:journals/corr/abs-2109-03154,
      author = {Bulathwela, Sahan and Yilmaz, Emine and Shawe-Taylor, John},
      title={PEEK: A Large Dataset of Learner Engagement with Educational Videos},
      booktitle = {4th workshop on Online Recommender Systems and User Modeling at ACM RecSys 2021},
      year = {2021}
  }
  

Change Log

This file documents all notable changes to the project.

• Date format is DD-MM-YYYY.

• TrueLearn adheres to Semantic Versioning.

• TrueLearn uses the Keep a Changelog format.

TrueLearn 1.1.0 (13-08-2023)

Added

• The option to visualise the variance in Bar, Dot, Line plotters. By default this is enabled except in the Line plotter.

Changed

• Interest Classifier now uses a binary skill representation.

• The INK Classifier can now be instantiated using a dictionary of parameters

• Added black border to Rose Plotter to make it easier to distinguish topics of similar colour.

• Locked the plotly, kaleido and wordcloud dependencies to avoid breaking changes.

• Radar plotter doesnt show the variance, due to concerns about the representation of variance being misleading.

Fixed

• Incorrect weights used in INK prediction

• Topic label overflow in Bubble plotter. This means that a reasonable number of topics (around 10 to 15) can be plotted for the default image size.

• Wordcloud now fully supports Python 3.11

Removed

• The orjson dependency, as JSON deserialisation is not our bottleneck and the additional dependency is not worth it.

TrueLearn 1.0.0 (14-04-2023)

• Initial release!

Authors and Contributors

KD-7, yuxqiu, sahanbull, deniselezi, aaneelshalman,