Automated Tutor Bot for Global Higher Education

Studious Digital Education are an award winning social enterprise who worked with UEA academics on an EIRA R&D project to develop a prototype of a web-based tutor bot to enhance digital learning.

 

Studious Digital Education are an award winning social enterprise founded by Dr Ivan Mitchell and James Gardiner. Studious have developed a forward-thinking digital platform called RYZE that inspires learning through content built from collaboration between academics, organisations and student partners.

The app offers students the opportunity to learn through interactive videos by academics and top organisations that explore a variety of educational subjects. Students can check their learning progress through interactive elements, such as quizzes and challenges, as well as sharing progress with friends and tutors.

Studious identified that the RYZE platform could fulfil a need within the Higher Education sector to develop digital learning resources in order to stay relevant and responsive in the digital age. After successfully applying for an EIRA R&D Grant, the team have been using machine learning systems to develop an automated ‘Tutorbot’.

 

The Challenge

 

The academic team behind the project included Professor Richard Harvey and Dr George Witty from the Schools of Computing Sciences and Mathematics at UEA. With EIRA funding support, the team built and developed a machine learning powered Tutorbot, which was designed to be assimilated into the Ryze Digital learning platform. The Tutorbot’s primary function was to automate content-based learning support, effectively emulating the role of an academic tutor.

The R&D project was developed before the global Covid-19 pandemic, but this past year has seen a step change in the way that learning has been delivered, with teaching moving online and hybrid models being employed by Higher Education Institutions (HEIs). The pandemic has further highlighted the relevance of a seamless transition to digital learning, and creating platforms that are intuitive and work to enhance students’ learning environment and experience.

 

The Approach

 

The objectives for this R&D project with Studious were:

  • to develop an interactive ‘TutorBot’ that facilitates complex interactions with students on subject-specific learning content using natural language processing
  • to advance the use of artificial intelligence and machine learning capability within the platform
  • to create a prototype framework for academics to develop and utilise at the subject level
  • to disseminate knowledge from the project to academic, teaching and EdTech practitioner audiences in the East of England.

To achieve these objectives, the project team was put together and included the academic expertise of Professor Richard Harvey, whose early career involved research into problems with signal processing, acoustics and vibration. When he joined UEA he was able to apply these interests to image processing and has been working in the fields of computer vision and artificial intelligence since then.

The project team further benefitted from the academic expertise of Dr George Witty, Lecturer in Mathematics and Widening Participation Academic Officer. After completing his PhD in mathematics at UEA in the area of representation theory, George was interested in using his mathematical backgrounds to explore the worlds of artificial intelligence and natural language processing.

“Working for Studious on an EIRA grant was an exciting experience for me. I had the opportunity to learn and develop my skills in areas such as artificial intelligence and computer programming, all whilst working on a project that was destined for a real app and its audience. This was quite a different challenge from my PhD, and it was interesting to see how my academic background could be used to support the project, whilst also learning new ideas to fill in the areas I had not explored before. The Studious team were a super friendly bunch and it was always fun having to work with the rest of the team on the many different elements of the project, such as how the chatbot should appear within in the app and what sort of questions the chatbot should answer. Despite working closely with a business, the exploratory nature of the project meant that I was still free to pursue the project in a way that best suited me, and this made it all the more enjoyable.” – Dr George Witty, UEA

The project started with an initial research phase, using design thinking and market research to define the problem. What was clear was that students want more access to clarification of their knowledge. Dr Witty then set to work understanding how to build a tutor bot which could get students get answers and clarify areas of uncertainty in their knowledge.

The team used the Blooms Taxonomy frame work, as a basis for understanding knowledge levels and types of knowledge acquisition.

Following research into different natural language processing services the team decided to use IBM Watson Assistant as the base of the tutor-bot project. The tutorbot was designed so that a student can ask a question, for example ‘what is leadership?’, and the tutorbot would then be able to give  a detailed description, offering opportunities to either analyse, evaluate or apply that understanding and creating an interactive dialogue.

 

The Outcome

 

The R&D project was successful in developing a web-based prototype of the tutorbot using dummy data. The system has also been integrated into the backend of the Ryze app for potential roll out in the future.

“It has been a privilege working with the EIRA team; the support from UEA has been fantastic and we’re delighted to have worked on this collaborative developmental project.  The pandemic has highlighted how critical digital technologies are for global education and we look forward to developing the next phase of this innovative work to support students in their learning across the world.” – Dr Ivan Mitchell, CEO Studious Digital Education

 

Next Steps

 

The next steps are to work on the implemented prototype  by using real data from a database. This is a challenging next phase because of the amount of content/data needed to be added. The business will seek further funding for this next phase of development.

The longer term aim is for the  developed tutorbot to give an additional personalised element to the RYZE app, which has potential to create added value. This additional functionality will be used to promote wider sales across the platform.