Twitter’s acquisition of Magic Pony (named after one investor’s attempt to analogise how unbelievable their technology is) has been one of the most high profile moves into machine learning so far.
As a result, we can reasonably expect our timelines to feature automatic sharpening of low-resolution and blurred video before too long.
But there are plenty of other less well-known applications that are doing equally cutting-edge things with machine learning.
Trooly is a Californian startup using it to automate part of the costly and labour-intensive process of running background checks on people. They announced their Series A just yesterday. Like us and our old Canary Wharf neighbours Digital Shadows, they apply their algorithms to digital footprint data – publicly available information by and/or about a person (or in our case, a company) – to make it easier and cheaper to assess risk and trustworthiness.
Meanwhile, IBM are using machine learning to predict the weather. Deep Thunder is ‘a machine learning-driven weather model developed by IBM Research to help industries ranging from aviation and agriculture to retail better predict the business impact of weather.’ IBM Watson is being used by business intelligence visualisation platform Buzzradar to gauge the emotion of an event, using a sample of tweets as training data, and just yesterday The Drum reported that Watson correctly predicted several Cannes Lions winners.
Zebra Medical Vision are using ML to analyse millions of imaging records and identify patients at risk of disease. Clinicians can then act to prevent the disease from ever occurring. Zebra just announced an extra $12m in funding.
Unfortunately, despite some frantic Googling I’m unable to find any evidence of a Russian startup our CEO Tom once came across that was using machine-learning in voice recognition systems installed at mineshaft entrances. Allegedly it was able to detect drunkenness in the voice and respond by keeping the mine locked down. Molodets.
Finally, there’s tronc. And no, I’ve no idea either.