When you’re talking to someone at a busty party, you can usually filter out the background noise and focus on the sound of the other person’s voice (unless it’s a very loud party, of course). Like a lot of human skills, it’s a remarkably hard one to replicate with technology – but two researchers reckon they has managed to do it using a ‘neural net’ that simulates a human brain. The only problem is that they aren’t sure exactly how the neural net is accomplishing the trick, as it doesn’t run on traditional algorithms.