AN SRUC researcher has developed a new computer tool to identify bovine tuberculosis (bTb) in dairy cattle which he is to share at an upcoming conference.

The annual NVIDIA GPU Technology Conference highlights breakthroughs in artificial intelligence, accelerated computing and intelligent networking and attracts around 100,000 attendees.

SRUC researcher Scott Denholm has been invited to give a presentation about how deep learning is being used in research to predict bTB – which affects around 37,000 animals and costs the country £176 million per year.

Mr Denholm's session will take place on Wednesday April 14 at 10am.

“Deep learning has enabled us, in partnership with National Milk Records, to develop a predictive tool capable of determining the TB status of individual dairy cows at an accuracy of as high as 99% and from as little as 30ml of milk," said Mr Denholm. "Milk samples are already collected as part of routine milk recording – a process practiced by around 60% of all dairy farms in the UK.

“This tool offers a completely non-invasive method to monitor dairy cows and alert the farmer of any animals likely to fail the main screening test for TB, enabling them to be isolated or removed earlier and restrict the spread of bTB within a herd.”

The conference, which is being held online between April 12 and 16, is free to attend and includes more than 1400 sessions covering the latest in AI, high performance computing, data science and more.

For more information visit https://www.nvidia.com/en-us/gtc/?ncid=GTCS21-NVMPALMER