Sometimes referred to as the father of modern computing, Alan Turing planted the roots of artificial intelligence and machine learning. The British mathematician deciphered cryptic messages during World War II, helping the Allies win. Yet, he was considered a criminal for his homosexuality. And because much of his work was classified, he did not receive due praise from his peers.

The vision

According to his biographer, Alan Turing envisioned machines “turned to any well-defined task by being supplied with the appropriate program.” The theories in Turing’s seminal paper, “On Computable Numbers, with an Application to the Entscheidungsproblem,” supported his invention of the universal Turing Machine, an abstract computing machine that provoked the creation of digital computers.

His digital program storage concepts set the framework for machine learning. Today, Turing’s ripple effect echoes through many industries including healthcare. Thanks to Turing’s first steps, computers categorize molecular data from biological samples to determine the root cause of disease.

The early days of machine learning

Turing’s design for the Automatic Computing Engine was based in his belief that humans and machines could think and learn similarly. He saw the cerebral cortex of an infant as an unorganized machine that became more organized through education. He noted that computing is learned and that both machines and humans could be trained to solve mathematical equations. He also thought that technology would one day embody intelligence in an artificial capacity.

Persecution

Turing’s greatest tragedy seemed to come just after he had reached acclaim. His wartime work to decode Nazi transmissions earned him an accolade as Officer of Most Excellent Order of the British Empire. In 1951 he was elected fellow of the Royal Society of London. But in 1952, Turing admitted he was in a relationship with a man. He was convicted of gross indecency and sentenced to 12 months of hormone therapy.

Turing was discovered dead in his bed, poisoned by cyanide in 1954. Officially ruled a suicide, his death is often attributed to an altered state of mind brought on by the hormone treatment. He had been publishing groundbreaking theories on patterns in living organisms — work still inspiring pattern recognition by machines today.

A continuing contribution

In 2009, Prime Minister Gordon Brown apologized for the British government’s treatment of Turing. In 2013, his work was declassified, and Queen Elizabeth II granted him a Royal Pardon. Turing’s theories were radical for his time but truly visionary.

Turing couldn’t have predicted the impact his theories on computing data would have. The Human Genome Project and the advanced computing abilities on which it relied applied Turing’s work in modern algorithms. Now, his concepts — far more than theories — help inform Data-Driven Medicine. To learn more about the patented algorithms of SOPHiA GENETICS applied to genomic and radiomic data, visit https://www.sophiagenetics.com/technology/.

At SOPHiA GENETICS, we believe that diversifying data modalities is the key to have better insights on diseases. That’s why, on top of genomics, our Radiomics R&D team is working on medical imagery analysis.

Applying their expertise to oncology specifically, members of the team just co-signed a newly published peer-reviewed paper hand-in-hand with several of our key partners in France. This promising paper, published in Neuro-Oncology, described the development of new evaluation criteria of treatment response in clinical trials for high-grade meningiomas.

Their observations and results support the need of the clinical community for new and improved evaluation criteria. Some criteria used in oncology today have not kept pace with the advances of precision medicine. Our team aimed to develop alternatives that have the potential to be better indicators of the response to current and future treatment protocols. Modern technologies such as AI, machine learning, and image segmentation can help develop imaging biomarkers that have the potential to redefine today's gold standards.

Thierry Colin, VP of Radiomics research at SOPHiA GENETICS and co-author of the paper, said: "We believe that radiomics capabilities can help better analyze clinical trials even in very complex situations and hope that it can be applied to everyday clinics as well in the future."

Thomas Graillon, neuro-surgeon at the Hospital La Timone in Marseille, France, and principal author, added: “Mathematical analysis performed by Thierry Colin’s team discriminated various response patterns to targeted therapies for aggressive meningiomas. Complementing the existing 6-month progression free survival method, this research enables new tools for the assessment of meningioma clinical trials and, in particular, the assessment of antitumoral drug activity in meningioma.”

Read the paper here.