AI changing the way scientists carry out experiments

October 22, 2018, Australian National University
Credit: CC0 Public Domain

There's plenty of speculation about what artificial intelligence, or AI, will look like in the future, but researchers from The Australian National University (ANU) are already harnessing its power.

The group from the ANU Department of Quantum Science has been experimenting with trapping at very cold temperatures, in their efforts to build a quantum communication network.

Now they've developed an with AI to help them run their experiments. Based loosely on the human brain, neural networks allow computers to "learn" to perform tasks.

"We use AI to control a large number of inputs to our experiment—the different laser and magnetic field settings—to seek out the best possible experimental conditions," said Dr. Geoff Campbell, post-doctoral fellow at the Centre for Quantum Computation and Communication Technology.

"Because we have so many inputs we can only make educated guesses based on our understanding of what works best, but the AI is better at it than we are."

Cold atoms are an important part of new technology like precision sensors and atomic clocks.

This latest research has demonstrated the potential for AI to optimise cold atomic systems. The solution found by the AI can trap twice as many cold atoms in half the time.

"The AI in our experiment found a solution that is highly effective and defies our intuition. One hundred students working for 100 years would probably never find it," said Associate Professor Ben Buchler.

"Our main research focus is to develop a quantum repeater, a device that can be used to send information over long distances. For that to work, we need to trap as many in it as possible. The difficulty is that it's almost impossible to build a complete model of how the atoms interact with each other, so we need to optimise the experiment by trial and error. The AI lets us do that more effectively."

While this AI was used for a specific task, the researchers are confident it could have much broader applications.

"AI is being used in an ever-increasing number of roles, from optimising the performance of our mobile devices to ," said Professor Lam.

"It's only going to become more commonplace as it gets developed into a general tool that can free up scientists from tedious optimisation of experimental settings in the lab."

The team believe the program could be used to improve a variety of experiments with a lot of input parameters, not just in the field.

Looking beyond pure research, they say it could also have a big impact on industrial and manufacturing processes by increasing efficiency and reducing cost.

"This is an example of a bigger trend where AI is becoming less exotic and soon will be seen as one more tool that the scientific community can use," said Professor Lam.

Explore further: Quantum computing a step closer to reality

Related Stories

Building 3-D atomic structures atom by atom using lasers

September 6, 2018

A team of researchers at Centre National de la Recherche Scientifique (CNRS) in France has developed a technique for arranging cold atoms into useful 3-D arrayed structures. In their paper published in the journal Nature, ...

Recommended for you

ATLAS experiment observes light scattering off light

March 20, 2019

Light-by-light scattering is a very rare phenomenon in which two photons interact, producing another pair of photons. This process was among the earliest predictions of quantum electrodynamics (QED), the quantum theory of ...

How heavy elements come about in the universe

March 19, 2019

Heavy elements are produced during stellar explosion or on the surfaces of neutron stars through the capture of hydrogen nuclei (protons). This occurs at extremely high temperatures, but at relatively low energies. An international ...

Trembling aspen leaves could save future Mars rovers

March 18, 2019

Researchers at the University of Warwick have been inspired by the unique movement of trembling aspen leaves, to devise an energy harvesting mechanism that could power weather sensors in hostile environments and could even ...

Quantum sensing method measures minuscule magnetic fields

March 15, 2019

A new way of measuring atomic-scale magnetic fields with great precision, not only up and down but sideways as well, has been developed by researchers at MIT. The new tool could be useful in applications as diverse as mapping ...

1 comment

Adjust slider to filter visible comments by rank

Display comments: newest first

not rated yet Oct 23, 2018
Question: is the experiment controlled in some way? For instance, if graduate students were using visualization software to watch over summaries of experimental runs in which automation widely varied the parameters in the same manner as the AI runs, would they now be able to compete against the AI? Or, put another way, is this study certain that the very same wide, automated ranges of parameters that were sampled in this AI experiment, have been attacked by humans using their visualization tools of choice?

Please sign in to add a comment. Registration is free, and takes less than a minute. Read more

Click here to reset your password.
Sign in to get notified via email when new comments are made.