Enhancing learning through Artificial Intelligence

Left to right: Dr. Rory Sie, Susan Israel and Debra Nolan, with Library chat bot student developers Joshua Cook, Wenjuan (Chloe) Sun,
Reece Cares Henriquez, Alexander Nicholson (behind) and Tianming Zhao.

How can we help students to learn as efficiently as possible and manage a large volume of information?

Artificial Intelligence (AI) in Education Senior Lecturer Dr Rory Sie, with UOW’s Learning, Teaching and Curriculum department, is implementing new technology to make the process of higher education study more accessible and supportive.

The idea to create AI in learning at UOW, was prompted by Dr Sie’s wife Elisabeth’s experience of studying in Australia. Being Dutch and having English as a second language, she found it difficult to learn, particularly with a lack of unified structure in learning material.

Elisabeth was struggling to cope with information overload and how to distil the high volume of resources available, while also being faced with an overwhelming number of assessments.

Her story reflects that of many students who are studying at UOW campuses in Australia, Dubai, Singapore and Hong Kong, and in many cases have English as a second language.

“The problem prompted the question; how can we solve this issue with AI? And wouldn’t it be great to have an intelligent assistant?” Dr Sie said.

“It’s a personalised assistant that really knows who you are, it takes into account your individual context, it knows that English is not your first language, it knows about your background and it has proactive intelligence - meaning it acts by itself rather than being reactive.”

Dr Sie is in the process of developing this intelligent assistant, currently known as ‘The Duck’. He explains how students could have an actual conversation with The Duck to receive personalised support in the form of recommending them with support services, remedial resources or peers that may help.

“It knows about the context of your situation, it knows if you are studying psychology for example, it recommends additional resources that are relevant at that specific moment in the session, it really interacts with you,” he said.

The Duck knows that Elisabeth is studying Research Methods and Statistics, and that she didn’t do HSC mathematics, and failed her formative assessment on calculating statistics by hand, therefore as part of its assistant abilities, it could recommend she read some additional course materials on this topic.

“It’s also important to distinguish that this technology is not designed to do things for students, but more so we are aiming to empower students to do things for themselves and develop the skills they need to learn - at least that’s my ultimate goal,” Dr Sie affirmed.

Artificial Intelligence works by creating fabricated neural networks by inputting large amounts of data into the computer and creating algorithms that categorise the data to help deliver relevant answers to students’ questions.

“I am working on a project with a group of students who are currently enrolled in their capstone project subject CSIT321 in the School of Computing and Information Technology. We are collaborating with the library to develop a ‘chat bot’ known as Moodji. We have a student who asks the librarian a question and the librarian processes that question and comes up with an answer.

“An artificial neural network takes all those words and processes them in what we call neurons - simulating the processing of information input into output - and then they find some answers,” he explained.

According to Dr Sie, as the volume of data and the computational power increases, the network of neurons is expanded to create more layers with more complex algorithms, enabling deep learning by the computer. The anticipated outcome is an ability to get more specific, intelligent and personalised answers to more questions.

Artificial Intelligence is not limited to just mapping questions to answers; it can map any input to an output. It is also used to predict future student outcomes (output) based on student demographics and behaviour (input) which can indicate whether they will pass or fail.

“We have the computer learn what leads to either a pass or fail. It makes a prediction about the student’s likely performance and if we know that, then we can create some interventions. And we could even have the computer learn about and prescribe interventions, which is already being done in healthcare at the moment.”

Dr Sie says the implementation of AI technology will not only make the learning process much easier for students, but it will facilitate their ability to get assistance around the clock, not just during office hours.