Renowned experts provide insights at Perse AI networking event
Two leading experts shared their wisdom on artificial intelligence (AI) in a special networking event at The Perse.
Dr Neil Lawrence, the inaugural DeepMind Professor of Machine Learning at Cambridge University, and Dr Anthony Cox, who led the expansion of AI and machine learning capabilities in the UK with llumina, spoke to a packed audience of parents, alumni and students in the Peter Hall Performing Arts Centre.
The event, organised by the Perse alumni and development team, began with an introduction from Head Ed Elliott, reflecting on the school’s success as a centre of excellence for computer science education, with 268 pupils currently studying the subject at A level and GCSE.
Outside The Perse, teachers and Upper Sixth students are helping pupils at Harston and Newton Primary School develop coding skills as part of the school’s community outreach programme.
More widely, the annual Perse Coding Team Challenge aims to test the programming skills of youngsters around the world, with around 7,500 pupils taking part in the competition last year.
Mr Elliott added that he was a “tech optimist” and used in the right way, felt AI could accelerate the rate of learning by speeding up research and improve the effectiveness of learning by giving students instant feedback on their work, as well as providing personalised learning tailored to individuals’ needs.
In his speech, Dr Lawrence, who is leading Cambridge’s flagship AI mission AI@Cam, put forward the notion of the ‘atomic human’ – the idea that each time machines are given the ability to do something, it cuts away at something previously unique to humans, and whether there comes a point where no more can be cut.
However, by cutting away at human intelligence that can be replaced by machines, he felt what makes us human in terms of vulnerabilities and limitations is uncovered.
Dr Lawrence went on to say that human judgements need to be made about technology to prevent a situation where we perceive it is always right, pointing out the faulty Horizon software system at the heart of the Post Office scandal.
However, used wisely, he added that AI could be used as a tool to enrich and empower people rather than humans becoming a tool of AI.
Meanwhile, Dr Cox, who is currently a visiting researcher at the Wellcome Sanger Institute, outlined how AI is helping gain a better understanding of the human genome.
He explained how predicting the shape of a protein from its sequence is a problem of huge practical importance, but that the use of AI system AlphaFold – “the ChatGPT moment for molecular biology” – had been deemed by many experts as solving the issue.
Dr Cox spoke about the 100,000 Genomes Project, the Genomics England initiative sequencing the according number of genomes from around 85,000 NHS patients affected by rare disease or cancer.
In terms of finding the cause of a rare genetic disease, he told the audience systems such as PrimateAI, which uses common variants of diseases in humans and other primates by training a neural network, which can be used to classify variants that have not been seen before.
He concluded that biological data is “noisy and uncertain” and it can be challenging to extract the data needed to train machine learning algorithms, but that AI had been used successfully to improve the ability to prioritise and interpret changes in DNA.