London scientists have created a machine that can repair its own damaged code
Good news for folks who regularly find themselves mashing control+alt+delete until their fingernails hurt: Researchers at the University College London have invented a new "systemic" computer that automatically repairs corrupted data whenever it appears. In non-geek speak, all that means is that this gilded, Platonic ideal of desktop PCs never, ever crashes.
How does such a miracle machine even work? Its creators, naturally, turned to Mother Nature — more specifically, the random modeling that allows swarms of bees to scatter in unison around a honey-hungry bear or that allows atoms to bounce around until they find their proper place.
"Today's computers work steadily through a list of instructions: One is fetched from the memory and executed, then the result of the computation is stashed in memory," explains Paul Marks at New Scientist. Crashes happen when a computer mangles these kinds of linear instructions and the code doesn't quite know what to do next. Hence, the dreaded error message. The new computer works a bit differently, dividing instructions into little digital slots called "systems." Here, an explanation:
Each system has a memory containing context-sensitive data that means it can only interact with other, similar systems. Rather than using a program counter, the systems are executed at times chosen by a pseudorandom number generator, designed to mimic nature's randomness. The systems carry out their instructions simultaneously, with no one system taking precedence over the others. [New Scientist]That means whenever a chunk of data hits a wall, these randomly-generated paths simply reroute these instructions through various systems (all containing the same message) to help it accomplish its original goal without shutting down. Kind of, but not exactly like how the human brain is capable of rewiring its neural pathways around obstructions. The creators of the device are optimistic, and think this intelligent, self-repairing tech could one day help damaged drones reprogram themselves on the fly, enable bioengineers to create more realistic models of the human brain, and trigger sentience in emerging mainframes like SkyNet. (One of these isn't for real.) The team plans on presenting their findings at a conference in Singapore this April.
Other stories from this topic:
- Instant Guide: Meet the spiky hedgehog rover that will bounce around Mars' moon
- Instant Guide: Meet Kenshiro: Japan's creepy boy robot with an aluminum skeleton and 160 muscles
- Essay: Falling in love with a bot
- Technology & Electronics
- University College London