Practice spaces
Practice spaces are a unique tool Examplary offers to make it easy for student to practice test questions for specific source materials.
In a Practice space, students can see their progress in mastering each of the topics set by the teacher, and get personalised feedback about their learning.
How practice spaces work
Selecting practice topics
Practice spaces are built around the source materials you upload, and the topics you choose. All questions generated will be directly related to a topic, and a student sees their progress by topic.
Infinite practice canvas
New questions are generated in the background constantly whilst students use the practice space. Based on the student's answers to questions, the space automatically selects the most appropriate next question to show them.
This works similar to your queue on Spotify or Apple Music: as the student works on questions, the AI will add additional questions to their queue that are relevant for them, based on their past performance on the specific topics, as well as various core learning principles.
The space will start out by asking questions to allow the AI to form a baseline evaluation of how well the student understands each of the topics selected for the practice space.
Spaced repetition and interweaving
Once a baseline is formed, the space will first start focussing on the topics the student needs the most help with, but interweave it with questions about the topics the student already has a good understanding of.
It also applies the principle of spaced repetition to foster long-term memory.
Personal feedback
Students get personal feedback after each question they answer. This feedback follows the following structure:
- What was wrong
- Why it was wrong
- What could be done to improve it
In addition, after each answer, overall and per-topic feedback for the student is generated. This feedback focusses on what they do well, and which topics and principles still need their attention. It aims to give them guidance on how to continue studying to achieve their goal.
Quality and safety
Ensuring high-quality questions
Questions generated in a practice space are always grounded in the source materials you upload to the space. This lessens the chance of hallucinations, and ensures the questions use terminology that matches what the student has read before and heard in the classroom.
In addition, each generated question is automatically scored using an algorithm that tests it against dozens of scientifically proven principles for good test questions. These include things like:
- Whether the question tests a singular concept
- Whether the question stem is a direct question or complete statement
- Whether the question language avoids unfairly disadvantaging particular groups
Questions that fail on a certain amount of these criteria are automatically dropped, and questions with high scores are more likely to be shown to students.
Teacher results dashboard
Teachers can at all times see student progress in the practice space. This includes a top-level overview of how the class is doing as a whole, as well as how each student is doing. It also allows the teacher to see all of the questions the student has answered, including their answers and the AI feedback they were shown.
Reporting mistakes and invalid content
Students are shown options to report incorrect or harmful AI-generated content. Teachers also have the ability to do this from their results dashboard.
When reporting content, a snapshot of that content and other non-PII context is shared with Examplary. This helps our team further improve the practice spaces system.
Harmful AI and student content scanning (coming soon)
Later this year, we'll be rolling out a system to automatically flag harmful content, both generated by AI, as well as in student answers. This will send alerts to the teacher, allowing them to take required safeguarding actions.