Building an ML Model for open-ended questions Answers

Question

Building an ML Model for open-ended questions Answers

Abdelmajeed Abdelmajeed 1

Hi All,

I am working on my thesis and require your help. Now I am trying to create text detector using Azur auto ML. the idea is the tutor will upload the model answers of open-ended questions such as ''What is Blockchain Technology?''. The model will the compare the student's answer with the model answers that is provided by the tutor. For example the model answer would be ''Blockchain defined: Blockchain is a shared, immutable ledger that facilitates the process of recording transactions and tracking assets in a business network. An asset can be tangible (a house, car, cash, land) or intangible (intellectual property, patents, copyrights, branding)''.
I do not how can I go about this model?

Thanks

LBM Solution 0 Reputation points

2025-06-27T12:31:56.91+00:00

To create a text detector using Azure AutoML for comparing student answers with model answers, start by collecting a dataset of model and student responses, labeling them as correct or incorrect. Preprocess the text by normalizing, tokenizing, and removing stop words. Use feature extraction methods like TF-IDF or BERT embeddings to convert text into machine-readable format. Upload the dataset to Azure, select text classification, and let AutoML choose the best model. Train and evaluate the model based on metrics like accuracy and precision. Deploy the model as an API for real-time comparison and provide feedback to students. You can find more resources on Azure AutoML here and text preprocessing techniques here.

1 answer

Your answer

LBM Solution 0 Reputation points

2025-06-27T12:31:56.91+00:00

To create a text detector using Azure AutoML for comparing student answers with model answers, start by collecting a dataset of model and student responses, labeling them as correct or incorrect. Preprocess the text by normalizing, tokenizing, and removing stop words. Use feature extraction methods like TF-IDF or BERT embeddings to convert text into machine-readable format. Upload the dataset to Azure, select text classification, and let AutoML choose the best model. Train and evaluate the model based on metrics like accuracy and precision. Deploy the model as an API for real-time comparison and provide feedback to students. You can find more resources on Azure AutoML here and text preprocessing techniques here.

Answer 1

Ramr-msft 17,826

@Abdelmajeed Abdelmajeed Thanks for the question. Here is the samples for nlp-recipes for similarity use case.
https://github.com/microsoft/nlp-recipes

Abdelmajeed Abdelmajeed 1 Reputation point

2022-07-21T14:17:19.57+00:00

I did not find the answer for my question here. May you help me further please?
Ramr-msft 17,826 Reputation points

2022-07-22T02:45:36.583+00:00

@Abdelmajeed Abdelmajeed Thanks, Currently Azure AutoML can't support this kind of feature, will be supported in the near future.

Can you please share the data and model that you have built.

Share via

Building an ML Model for open-ended questions Answers

1 answer

Your answer