Overview / Usage

It is very difficult for a student to do self-assessment without any personal tutor or examiner. Our system wants to remove this student/tutor dependency by providing relevant questions and then when a student enters the answers our system analyses the input and provide the proper report with which a student can understand his/her weak and strong points.

On the other hand, our system also generates the quick questions for the given topic or content, hence this system can also be used by the Professors, Researchers, Tuition classes and many other educational institutes for a quick and accurate question generation.

Moreover, the question generated by our system is completely unpredictable priory as it is computer generated and do not base on the mind of the particular person. Hence all students, teachers, professors, researchers, analysts etc. those who want to do self-assessment can use this system for the self- evaluation. Tuition classes and other educational institutes can use this system for question paper formation.

If someone wants to do self-assessment or want to prepare quick examination for someone else then at this current moment we have to do it in a traditional way of depending on the tutor or mentor for the preparation of such examination. This way is traditional and very time consuming moreover we have to remain dependent on someone for our own preparation. We want to remove this dependency by making a system which can generate the quick examination for you from the content provided by you and then it must also take input from the user and then it should analyze the input with the actual answer.

Until this moment we were always dependent on someone else for our own assessment and if we want to generate a quick examination paper set then we have to contact the experts for that. This way of solving the same problem is very traditional and very much time-consuming. This is the problem which we can automate by using Artificial Intelligence and Machine Learning by the use of Natural Language Processing.

Methodology / Approach

Our system works with the following strategy, initially, the first step selects the best potential set of sentences from the given text input from which we can possibly generate the questions. Next step(second step) is to find the subject and context of the sentence. In the third step, we will try to analyze which is best the form of a question that can be generated from that sentence.

For this process, we will use Encoder-Decoder nets that have been used by Google for it's Natural Machine Translation (Language Translation) and Recurrent Neural Nets. By keeping Encoder-Decoder at the core, we also take help from Stanford Parser and NLTK for grammar analysis and more basic natural language analysis. The whole programming is done with the help of Intel Optimised Python distribution which makes the working very fast and efficient. We also use Intel's Devcloud for testing our data models.

Technologies Used

Intel Optimized Python Library, NLTK, Stanford Parser, Keras, Tensorflow

Repository

https://github.com/adityasarvaiya/Automatic_Question_Generation

Other links

• Blog on this Project