Babe, good afternoon! I just rushed home as soon as I got off work because my good friend said she made delicious hairy crabs for me. I ate two of them in one go, they were so tasty!

During a casual chat, we happened to discuss the education issue. To be honest, I've noticed a problem when using educational apps.
Either the questions are all standardized content that is monotonous and do not match one's own learning progress.

Sometimes I encounter a few customized questions, and it always feels a bit flawed. It could be that the knowledge points are off-target, or perhaps the answers are not rigorous enough.

Later I realized that doing personalized education assessments is really not as simple as just calling an AI model.

The team at Learnrite specializes in this area; they do not want a one-size-fits-all test, but rather aim to tailor questions for each student.

But it got stuck at the very beginning; high-quality educational questions are too difficult to create because they need to combine the details of subject knowledge and comply with the teaching syllabus. Previously, the cost of each question was as high as 5 dollars.

The more troubling part is that when using those popular AI models to generate questions, the accuracy rate for medium-difficulty questions is only 75%. There are too many mistakes, and in the end, we still have to rely on humans to review them one by one, which is costly and time-consuming, completely failing to showcase the advantages of AI.

It wasn't until they used @miranetwork's tools that they solved this problem.

Mira has a "verification generation" feature that not only creates questions but also automatically checks accuracy, improving the accuracy rate to 96%. There's almost no need for manual review anymore, and the question generation speed has decreased from one question per hour to just a few minutes, significantly lowering costs.

To be honest, I think this is how AI should be - not just pursuing what it can do, but solving real problems.

In the past, many educational apps struggled to achieve personalization because they had to ensure quality while controlling costs, making it impossible to strike a balance.

Mira helps them bridge this critical link, allowing students to use questions that match their own level, while institutions can operate on a larger scale. It can be said that it's a win-win situation!
#Mira