here you go:AI chipping away at human recognitionBrainchip’s...

here you go:

AI chipping away at human recognitionBrainchip’s Studio software scanning faces.You may have watched episodes of CSI or BlackMirror where a computer flashes through thousands of images toidentify a crime suspect in seconds. If so you’ll understand some of the workof BrainChip, a San Francisco-based company that is listed on the AustralianSecurities Exchange with $40 million in local investments.BrainChip has offered this solution for more than a year throughsoftware called Studio and a hardware accelerator. It’s currently creating thetechnology on a chip. Security cameras record vision that is analysed toidentify people and objects, often in real time. These systems are advanced,especially in countries such as China.Chief executive Louis DiNardo says its vision system is already workingat several places in France, such as the Paris Metro, the municipal policeforce in Paris, France’s equivalent to a department of homeland security, andBordeaux airport. He says trials are under way in Spain, Britain, Sweden,Norway, Denmark, and the US. New systems should operate early next year.its artificial intelligence to detect counterfeit chips at casinos.Cameras glean the chips being played, and the system tests their authenticity.The system is also being adapted to help Safran, the French aerospace company,to verify the correct installations of equipment during the assembly of Airbusplanes. A device flashes green if a wire installation is going into the correctsocket, or red and sounds an alarm if assembly is incorrect.DiNardo says the human recognition system goes beyond conventional imageAI. It works more like the human brain, generating “bioelectrical spikes”: otherprompts apart from physical appearance that can identify someone. They callthis “neuromorphic computing”. DiNardo says this approach drastically reducesmachine learning time. BrainChip is also launching Akida, a system-on-a-chipthat DiNardo says is the company’s future. Akida is not confined to visionanalysis; to uncover malware it recognises changes to patterns such as packetsof data flowing into a network. It can also be adapted to recognise tradingpatterns in finance. In Australia, it is earmarked for use in agriculture.