Researchers continue to run at the second hurdle on the race to AGI.
Researchers at DeepMind have built an algorithm to mimic the way in which infants learn the basic rules of physics. One such would be that a solid object cannot pass through another solid object. The algorithm, dubbed PLATO, is rather successful. The conclusion, though, is perhaps not so surprising. A bit of prior knowledge makes a big difference to the success of the learning effort. The reason this type of research is so important is that there are only two hurdles left in the race to develop artificial superintelligence. The first hurdle is to instantiate consciousness in a robot. The second hurdle is to recreate the flexibility of human thought process in AI. DeepMind’s research is squarely aimed at the second hurdle. Check out my book to see how we might tackle the first hurdle.
Link to article: https://www.nature.com/articles/d41586-022-01921-7
You may also like to browse other AI posts: https://www.thesentientrobot.com/category/ai/