Eedi.com
Advancing pedagogy responsibly using big data to power our mission to help every child build the skills and confidence they need to thrive in a rapidly changing world.
About us
Eedi hosts the largest repository of diagnostic questions in the world. With over 120 million answers we have a unique data set with which to understand students’ misconceptions. In collaboration with Microsoft Research we have developed ground-breaking machine learning models to learn from this data and understand and direct students’ learning journeys to help improve learning outcomes. We are committed to sharing our findings and data to further the field.
Opportunities
Join the new Machine Learning team!
Deadline: 1st Dec 2022
Competitions
In 2020 we hosted the NeurIPS Education Challenge to predict student responses, determine question quality, and identify a personalised set of questions for each student. Our competition received 3696 entries from 382 teams, and many research papers have been published using the competition data set. In 2021 the data set that we released won the Best Publicly Available Educational Data Set Prize awarded by the Educational Data Mining Society. We continue to support the data set as part of our ongoing learning engineering project.
We ran a follow-up challenge at NeurIPS in 2022 which involved the discovery of causal relationships between constructs and inference on the impact of learning one construct on student's success in learning another. The dataset we released for includes real-world AB experiment data which can be used for evaluating causal models.
We ran a follow-up challenge at NeurIPS in 2022 which involved the discovery of causal relationships between constructs and inference on the impact of learning one construct on student's success in learning another. The dataset we released for includes real-world AB experiment data which can be used for evaluating causal models.
Experiments
We're opening our lab doors! Join us as we explore ways to improve student engagement.
Read our blog about the first open A/B experiment dataset here.
Read our blog about the first open A/B experiment dataset here.
Our team
Simon Woodhead
Chief data scientist and co-founder
Digory Smith
Data scientist
Craig Barton
Teacher, best-selling author and co-founder
Rachel Kidson
Teacher and Educational Psychology Lead
Panagiota Konstantinou
Machine Learning Research Engineer
Bibi Groot
Director of Learning Sciences
Publications
CausalEdu: a real-world education dataset for temporal causal
discovery and inference
CLeaR 2023 Datasets Track 1–9, 2023
Simultaneous Missing Value Imputation and Structure Learning with Groups
https://arxiv.org/abs/2110.08223
CoRGi: Content-Rich Graph Neural Networks with Attention.
In Proceedings of the 28th ACM SIGKDD Conference on Knowledge Discovery and Data Mining (pp. 773-783).
NeurIPS Competition Instructions and Guide: Causal Insights for Learning Paths in Education
arXiv preprint arXiv:2208.12610
VICause: Simultaneous Missing Value Imputation and Causal Discovery with Groups
arXiv preprint arXiv:2110.08223
School students’ confidence when answering diagnostic questions online
Educational Studies in Mathematics, 1-31
Contextual HyperNetworks for Novel Feature Adaptation
arXiv preprint arXiv:2104.05860
Results and Insights from Diagnostic Questions: The NeurIPS 2020 Education Challenge
In NeurIPS 2020 Competition and Demonstration Track (pp. 191-205). PMLR
Diagnostic questions: The neurips 2020 education challenge
arXiv preprint arXiv:2007.12061
Large-scale educational question analysis with partial variational auto-encoders
arXiv preprint arXiv:2003.05980
On Formative Assessment in Math: How Diagnostic Questions Can Help
American Educator, 42 (2), 33.