Tutor HuntBlog

Machine learning - what can it offer the education sector?

Schools
all2

Someone hearing about machine learning for the first time might imagine a robot at the front of a classroom, delivering the lesson in a tinny voice. Machine leaning is actually a complex software algorithm that analyses patterns of behaviour, and makes predictions using this data. Think of the way Netflix or Amazon makes `suggestions` about what you might like to watch or buy - thats all done with Machine learning.

Something so powerful is inevitably going to have applications in many different fields, and the company Better Examinations has been using machine learning to supply services to the education sector. Founded in 2012 in Dublin, the company has seen a surge in demand during the pandemic, and has even had to scale up its operations to meet demands.

The Tech company, owned by Piero Tintori, has developed software that allows thousands of students to sit virtual exams. Wherever they may be located, at home, or in a scholastic on-site bubble, the student can use their laptop to log into the exam portal. This in itself is nothing particularly novel, as there are an abundance of existing software platforms that can perform such tasks - what makes Better Examinations` product unique is that it can detect if a student is attempting to cheat through the procedures of machine learning.

Setting exams in anywhere but an exam hall has always proved problematic: without the ever pacing watchful invigilator the opportunities to cheat are so abundant, even the most honest student might succumb to checking the internet for that elusive answer. Better Examinations seems to have solved this problem by using the very latest machine learning (ML) technology, which can be considered to be a particularly advanced form of artificial intelligence.

Their software can `detect patterns` in an examinee`s behaviour that suggest they are cheating. What exactly these patterns are remains for the moment rather nebulous, but you can understand the company would not wish to disclose the confidential workings of their product, especially if, by doing so, they might undermine the efficacy of the software. One impressive feature of the technology is that it can perform a visual identity check on the examinee, ensuring a precocious friend or well meaning parent isn`t sitting the test for them.  Using their computer`s webcam, the software performs a facial recognition test, checking the live image against an uploaded photo that has been approved by a teacher. The software will perform these visual checks periodically, throughout the exam, to ensure the examinee does not trade places with someone else. It also restricts access to the internet, and any applications on the computer that could be used to help the student.

The software, which has been available for a few years now, has understandably been in huge demand during 2020. With the pandemic leading to to lockdowns and closures of schools and exam halls, thousands of students have been unable to take assessments under exam conditions.

`We had 60 organisations from all over the world contact us out of the blue,` Said Mr Tintori, `who wanted to run exams online in May and June, everything from universities, to professional organisations, to schools.`

The machine leaning used by Better Examinations is being put to use in numerous sectors, and we may well be in the early stages of whole new methodology of data analysis. A major UK retail bank has declared that it has recently started using ML algorithms, in order to identify customers whose pattern of transactions may identify them as being in financial difficulty. The bank automatically contacts these account holders, and provides them with a support package, the aim being to offer preemptive help before matters become more serious.

BJSS, the multinational technology company providing this ML software to the banking sector, has stated that it is cost effective to identify customers who are facing financial hardship as early as possible. During their initial assessment of the retail banking sector, they found that of those in arrears, more than 30% were in such a precarious position that `there was very little the bank could do for them at that point.`

One can only imagine what benefits machine learning will provide the education sector in the future. Perhaps by analysing their past assessments it will be possible to specifically tailor online lessons for each pupil, the bespoke classes providing a carefully constructed pedagogy, which will ensure they are taught in such a way that their individual strengths and weaknesses are engendered and attended to.

3 years ago
comment