Natural Language Processing Functionality for a North-American E-Learning Technology Vendor

Challenge

A leading North-American e-Learning technology vendor, engaged in the development and marketing of a suite of highly innovative proprietary products that support e-learning activities in more than 100 languages, approached Envion Software for assistance.

The company was interested in our long-standing business domain expertise in NLP, and sought to employ this expertise to develop a speech pattern recognition and analysis module for one of their flagship products.

The client’s system is a major language-learning framework geared toward foreign language learners. The solution supports multiple learning modes, such as, for example, writing exercises, guesswork, pronunciation exercises, and more.

The functionality Envion was requested to develop is designed to provide unsupervised pronunciation training to language learners. It is also used to evaluate with precision their pronunciation skills.

After a language learner reads aloud a phrase displayed by the system, the application displays a wave form, pointing out and detailing their mistakes, if any. This is done by the system by highlighting the corresponding word, syllable, or phoneme, and providing the correct pronunciation. In addition, the application displays a prompt that contains any other existing pronunciation options.

In order to evaluate a language learner’s pronunciation skills, a diverse array of sophisticated technologies and techniques are used, including artificial neural networks, statistical learning, machine learning, decision trees, deep learning, pattern matching, and more.

 

Solution

The project has been ongoing for 4 years now, and Envion has already delivered all the requested functionality that is, currently, only being expanded by our project team.

The project team communicated directly with one of the client’s subject matter experts to get a better idea about the client’s requirements for the functionality to be developed.

The amount of research the Envion team needed to conduct in order to encompass all the client requirements, as well as the more optimal implementation of these requirements, constituted the biggest challenges, posed by the project.

 

Technology Stack

The Envion team used the following technologies to implement the project:

  • Java8
  • HTML5/CSS3
  • Python
  • MongoDB
  • GlusterFS
  • C++
  • Matlab
  • RabbitMQ
  • Redis
  • Tomcat
  • Gunicorn

Result

Relying on Envion’s best-in-breed expertise and broad experience in developing e-learning and speech processing solutions, the client has received state-of-the-art, mission-critical  functionality that allows them to market their now complete and fully functional solution a lot more gainfully.

Contact Us

captcha