Innovation Takes a Long Time

One of the counter-intuitive aspects of technological developments is that they don't happen suddenly. The public doesn't hear of a new product until it is ready for sale, not the decades of development which has taken place. But the success of these products is not inevitable and even experts can't tell what will work and will succeed in the market and what will not.

Back in 2014, I volunteered to judge a Random Hacks of Kindness hackerthon hosted at the Commonwealth Bank headquarters in Sydney. One of the entries was a data analysis system for the BlueBottle USV from Ocius. This is a small solar powered vessel, for collecting oceanographic data. We awarded the prize for the competition to another entry, which shows how little we knew. Military drones are now a hot product, and Ocius Technology have gone on to successfully sell their USVs for military survalence.

At the time the Australian Minister for Defence criticized the Australian shipbuilding industry saying they did not trust the government owned shipbuilder to "build a canoe". The irony here is that the BlueBottle looks like a canoe and its military value in not being anything like the large expensive warships favoured by government. Another irony is that the Australian Government damaged the local shipbuilding industry, and weakened national defence, by chopping and changing decisions to purchase major warships. The government seemed to think you could pick a warship off the shelf like an appliance, when this involves decades of work. A third irony is that indecision may turn out to be useful, as the delayed major procurement can be replaced with smaller drones.

Calibrating AI For Saftey

Greetings from the Australian National University, where I attended the weekly AI, ML and Friends seminar. Jiawei Liu was speaking on "Uncertainty Calibration for Deep Neural Networks". I didn't understand the equations, but the idea, as I understand it, is to have a measure of how well AI is doing.

What got my attention was an example of recognising a speedboat. From the original photo it was clear to a human what this was. But the black and white outline, as used by some algorithms, looked to me like a USV (Uncrewed  Surface Vessel). These are being used, right now, in the Black and Read Seas, laden with explosives, to attack ships. The crews of warships have to maintain a lookout 24 hours a day, for weeks, trying to spot drone attacks. The USVs are modified speedboats so it is difficult to tell them from fishing boats. This type of AI analysis can help ensure mistakes are not made which could result in the loss of a ship, or the sinking of an innocent fishing boat.