It was just a few years ago when the scientists from Google started working with the neuroscientist Steve Finkbeiner team to make advances in scientific discovery with the help of Google technologies. They were interested in using the deep-learning approaches to mountains of imaging data created by the Finkbeiner team. Deep-learning algorithms fetches raw features from an exceedingly large data set and then, use it to make a predictive tool with the patterns buried inside. The algorithms can later be used to analyse other data too.
This can help tackle really tough set of problems and see structure in large amounts of data. Finkbeiner's team has produced a lot of data through the high-throughput imaging strategy called robotic microscopy. But they were not able to analyse the data at the speed it acquired them. So, they were happy to collaborate. Those efforts are significantly paying off now. Deep learning that is one of the most promising branch of artificial intelligence is paving its way into biology. These algorithms have already penetrated our modern life in smartphones, speakers and self-driving cars. In biology, deep-learning algorithms are able to dive into the data in ways humans cannot and thus, help in detecting features that are difficult to catch. Researchers are using these algorithms to differentiate cellular images, for advancements in drug discovery, make genomic connections and discover links between different data types from electronic medical records to genomics and imaging.
By: Neha Maheshwari