In February 2021, the archive covered research led by MIT’s Caroline Uhler into identifying existing drugs for further investigation against Covid-19. The work considered gene-expression changes associated with both infection and ageing.
The approach combined an autoencoder with analysis of interacting genes and proteins. It used those relationships to prioritise possible targets and candidates for further study. The original MIT coverage identified RIPK1 as one target of interest.
The researchers explicitly distinguished computational prioritisation from clinical evidence. Finding a candidate does not establish that it treats Covid-19 safely or effectively. The archive records an application of machine learning to research, not treatment advice. Its central idea was to help researchers decide what to investigate next, with efficacy remaining a question for subsequent testing.
Our October 2022 industry overview looked at five areas exploring machine learning: healthcare, transport, finance, retail and property valuation. Each used prediction or pattern recognition, but with very different data and consequences for mistakes.
The October 2022 archive explored how computer-aided systems could support clinicians by organising images, records and monitoring data. Its central theme was assistance with information-heavy work, with diagnosis and treatment remaining clinical responsibilities.
In March 2021, researchers at UC San Diego reported a device designed to perform a neural-network activation function directly in hardware. The work explored a way to reduce the circuitry and energy involved in that part of a network.
Tell us about your project — we'll tell you honestly whether we can help.