Deepwise

Deepwise

E- learning environment that adapts the abilities of any student.
Challenge

Challenge

The challenge of this project, co-financed by the Ministry of Energy, Tourism and Digital Agenda. (Num. TSI-100105-2017-9), has been to create a student-machine interaction that had into account the variety of educational resources, created by the teacher and the student´s interests and his/her intellectual abilities.

This R&D project aims to develop an intelligent platform that, through BigData, Machine Learning and Natural Language Processing techniques, can adapt to the specific needs of each student.

Adaptative Learning

Adaptative Learning

Adaptive learning implies a conception of each student as a unique individual with a particular type of intelligence. Each one will have their rhythm of studies and will use different ways to achieve their goals.

Emphasis arranged on the importance of addressing the strengths and weaknesses of each student while adapting the educational system to the needs of each student.

Our solution

"Creating the ML platform, for adaptive learning / OpenEDx
A technological and user experience challenge."
Data research
Data research
Thanks to Data Mining, a personalised offer of courses are offered, based on the student's profile and interests; and with Machine Learning, the personalisation of the courses is sought, adapting the itinerary and the syllabus by categorising the student according to Gardner's intelligence.
Language Processing
Language Processing
It is possible to infer knowledge from the information gathered through Big Data and create educational agents capable of responding to the natural language of students in forums and virtual assistants, through an intelligent and automatic response, based on the learning of previous editions.
Machine learning challenges
Machine learning challenges
Indexing of the intelligence, learning styles and students interests employing self-study algorithms that allow that the recommendations made by the adaptive learning module develop and improve over time. The algorithms suggest the recommended courses and contents to each student in the platform based on the current alternatives and based on their intelligence type.

State of the arte enhancements

Categorization of the intelligences, learning styles and interests of the students, through self-learning algorithms. The algorithms propose to the platform the recommendations of courses and contents to each student, based on the existing alternatives and their type of intelligence.
Results

Results

The solution based on different components:

  • MOOC e-learning platform
  • Big Data Module
  • Natural Language Processing Engine
  • Adaptive Pedagogic Agent Module

Technologies

Talend
Hadoop
Mongodb
Hortonworks
Spark
Solr
Elastic
Apache Hbase
Hive
Kafka
Mahout
Mapreduce

The project has been co-financed by the Ministry of Energy, Tourism and Digital Agenda.

Project number: TSI-100105-2017-9
We want to help you achieve your digital objectives. Let's talk!
base pixel px
Convert
Enter PX px
or
Enter EM em
Result