Frequent data collection on the size of individual animals can provide important information for evaluating growth rates and size distributions, which provide insights into productivity, conditions of the pond and potential yield. This information can help the farm manager predict and avoid unwanted situations. Prawn farm technicians pull up feed trays as part of their daily workflow to understand feed consumption and adjust feed rates. The tray typically captures a good number of prawns because feed is added to the tray to attract the prawns. We aim to take advantage of this practice as this process is more frequent (once/twice daily) than the casting of a net (once every week or fortnight).

Both Microsoft and Google envision a not-so-distant future in which donning smart headgear to repair an elevator or assemble a tractor motor is the norm. Over the past several days, both tech giants have revealed they're each taking concrete and compelling steps to make that happen. Microsoft just revealed that it's working on a new artificial intelligence chip to power its second-generation HoloLens headset. The coprocessor's chief job will be implementing deep neural networks -- a machine learning technique with a structure that loosely resembles the human brain -- into the HoloLens' core processing unit. A dedicated A.I. chip is necessary for a gadget like the HoloLens, says Microsoft, which must be able to comprehend large amounts of complex data gathered by its depth and camera sensors without latency.