top of page

Symposium Series on "DataScience,  AI & ML

A very informative and insightful symposium was organized by School of Technology, GSFC University on the emerging area of AI, Data Science & ML on 24th Oct, 2020 at 10 am. Speaker of the symposium was Dr. Ankit Desai, Principal Data Scientist from Embibe, Bengaluru.


Embibe is an education platform, which leverages data analytics to deliver personalized learning outcomes for students. With AI stacks focused on content intelligence and automation, behavioral recommendations, and student intelligence, Embibe’s products are redefining the way EdTech can impact the lives of students and teachers. Reliance Industries limited has invested in Embibe to form India’s Largest AI based education platform.


Dr. Ankit Desai gave a very brief idea about what are the goals of Embibe and the emerging filed in which Embibe is concentrating.

Key Takeaways with the glimpse of the symposium are:

  1. The Future of EdTech: An AI platform for Education

  2. Vision of improve the learning through online platform

  3. As recently presented in 43rd Annual General meeting (AGM) 2020 of reliance industries, Three foundational blocks are:

    • Unprecedented Personalization – Using knowledge graph, Student can go back to standard to solve and clear the building concept of any subject.

    • Incredible Content – Assembling world’s best content from different resources, Comprehensive question bank, Student experience based analysis to create better learning platform.

    • Teacher Empowerment – Help government to reach more students with higher quality of education, reliable delivery and illustrate art of the platform with personalized homework and high quality test

  4. Important pillars of Online Education which converts education to Educational Intelligence,

  5. Major components to build AI Based education platform.

  6. Content Building through knowledge graph which will give the relationship between the concepts without focusing on some particular board learning. The main aim of the content creation is to prepare higher order content which emphasis on creative learning not on rot learning. Through knowledge graph student can identify and solve the concept where student need revision and solve their queries. There are some competency in the knowledge graph in every concept to check weather student knows the particular concept. This module is based on semi supervised machine learning where it can read the digital book and map the concepts automatically.

  7. Deep learning based automated module can read the unstructured data and identify the sentences eligible for fill in the blanks and also identifies the filler of the blank.

  8. Distractor generation for MCQ which improves the difficulty level of the question. More close options will increase the difficulty level of the MCQ. This module is based on the concepts of Deep learning.

  9. Translator module provided by the Embibe is approximately 35% better than the Google translator in the academic domain of translation.

  10. Smart tagging module can understand the question and images of the question and can tag that question with the concepts. It also identifies the difficulty level and time to solve the question.

  11. Smart Book module parse the entire digital book and constructs the list of concepts used in particular chapter and which concept was discussed in earlier chapters. It automatically generates the description of the concept in easy language. It suggest best digital videos from different resources with particular concept itself. Also display’s the relevant images better than the Google images. The module named Coobo identifies student is engaged during the tutorial or video by capturing all the signals from learner. Normally students highlights or underlines the key words in the books based on the importance of that word in the particular topic. Similarly deep leaning based module identifies the word in essential, importance and primary category.

  12. This modules also prepare the questions by parsing the text which has different difficulty level and also extract the back of the chapter questions from every book with its translation which reduces the human efforts for preparing different level of question papers.

  13. Embibe will introduce one more module based on Optical Character Recognition which can read the text of your screen shot of your handwriting in any language. Not only reading but this module will understand and solve the question with step by step solution for better understanding.

  14. Automatic test generation module reduces the teacher’s effort as well as prepare same difficulty level paper sets for the same examination with equal weightage to every chapters. This module also display’s the test quality score. This algorithm is already used by Government of India to prepare competitive examination papers.

  15. Embibe had proved that qualifies the test not only based on the domain knowledge but also based on the behavior how you give the test.

  16. At the end Dr. Desai had discussed the patent published by the Embibe for different technologies and their awards received and also give opportunity to the students for an internship in Embibe.

  17. Students were very excited and asked lots of queries to the expert. Students were very curious about the working of different modules. It was very interacting symposium and concluded with the goal of meeting again for some workshop or internships.

Feedback of student: Feedback of students were very satisfactory and expert had also given good feedback about involvement of students during the symposium.


Recording of the Session:

Pillars of Online Education
Knowledge Graph
OCR tool
Major components to build AI Based educa
MCQ Generation
Expert of the symposium
Example Solution module
Deep Learning based automatic  fill in t
Collaboration with Reliance
Behaviour analysis
Automated Translation tool
bottom of page