The extent to which classmates and teachers were investing in cryptocurrencies prompted Taha Fareed and Jevin Joy from Coláiste Phádraig in Lucan, Co Dublin, to develop their own computer model using artificial intelligence to predict future cryptocurrency values.
Having taught themselves about machine learning and algorithms with the help of video tutorials, the 15-year-olds are about to launch a website where the public can use their model to predict the course of their investments.
Cryptocurrencies, which allow payments to be made electronically, may not yet be the dominant form of financial transaction, but this may soon change as they can be safer in avoiding fraud, according to Jevin – although people are often scared by their fluctuations in value.
The students’ model, which evaluates previous data to predict future valuations, has notable precision, Jevin added. This is indicated by “low losses”, which confirm reliability in a world where fluctuations are the norm.
Their method uses Bayes’ theorem and Bayesian optimisation to enhance AI, and is among the projects in the running at this year’s BT Young Scientist and Technology Exhibition (BTYSTE), which began on Wednesday. “It’s incredibly good and better [than] the best model we have come across in our research,” Taha adds of their method. “We’re very happy with what we have achieved but we know there are a lot of further improvements we can make.”
Young Scientist veterans
The third-year secondary school students are BTYSTE veterans, with this year being the third time they have entered. Overall winners of the contest often enter over a number of years in their preferred area of research before winning top honours. Remarkably, however, under the guidance of their computer science teacher Kevin Carey, the students have had entries on three different topics.
Their first was on using computer technology to help people on the autism spectrum to reduce their sensitivity to the brightness of lights and the volume of sounds. Meanwhile, last year’s entry was on software deploying game theory to improve strategy in a computer game, for which they won an award. This year they hope to be in the winners’ enclosure again.