Short Bytes: Using artificial intelligence, physicists at Australian national university have been able to mimic a Nobel-winning experiment i.e. creation of a Bose-Einstein condensate. This AI took around a dozen of repetitions to learn and reproduce the experiment result.
The invention of such AI machine would facilitate physicists to focus more on the high-level problem rather than focussing more on the abstract levels.
The experiment performed by the AI was related to the creation of a Bose-Einstein condensate. Bose-Einstein condensate is a hyper-cold gas for which three physicists had won the Nobel Prize in 2001.
Australian physicists and AI re-created the experiment something like this:
- First, they cooled down the gas to 1 microkelvin and then the control was passed over to the AI machine.
- AI, through its learning, was successfully able to figure out ways to apply lasers and also control other parameters side by side to cool down it to a few hundred nanokelvin.
- This feat was achieved after almost tens of repetitions. With each repetition, the AI got smarter and was more efficient in finding out the perfect way through learning.
Paul Wigley, a co-lead researcher says about the research:
AI could be applied more in other physics automation experimentations where learning through repetition is needed.