Robot Makes Scientific Discovery All by Itself

For the first time, a robotic system has made a novel scientific discovery with virtually no human intellectual input. Scientists designed “Adam” to carry out the entire scientific process on its own: formulating hypotheses, designing and running experiments, analyzing data, and deciding which experiments to run next. “It’s a major advance,” says David Waltz of […]

Adam

For the first time, a robotic system has made a novel scientific discovery with virtually no human intellectual input.

Scientists designed "Adam" to carry out the entire scientific process on its own: formulating hypotheses, designing and running experiments, analyzing data, and deciding which experiments to run next.

"It's a major advance," says David Waltz of the Center for Computational Learning Systems at Columbia University. "Science is being done here in a way that incorporates artificial intelligence. It's automating a part of the scientific process that hasn't been automated in the past."

The demonstration of autonomous science breaks major ground. Researchers have been automating portions of the scientific process for decades, using robotic laboratory instruments to screen for drugs and sequence genomes, but humans are usually responsible for forming the hypotheses and designing the experiments themselves. After the experiments are complete, the humans must exert themselves again to draw conclusions.

Meanwhile, some software programs can analyze data to generate hypotheses or conclusions, but they don't interact with the physical realm. Adam is the first automated system to complete the cycle from hypothesis, to experiment, to reformulated hypothesis without human intervention.

Adam's British designers, led by Ross King at Aberystwyth University in Wales, acknowledge that the robot's discoveries have been "of a modest kind" thus far. Its proving ground as a scientist has been the genome of baker's yeast, a popular laboratory species. Baker's yeast is one of the best understood organisms, but 10 to 15 percent of its roughly 6,000 genes have unknown functions. The scientists hoped Adam could shed light on some of these mystery genes.

They armed Adam with a model of yeast metabolism and a database of genes and proteins involved in metabolism in other species. Then they set the mechanical beast loose, only intervening to remove waste or replace consumed solutions. The results appear Thursday in Science.

Adam sought out gaps in the metabolism model, specifically orphan enzymes, which scientists think exist, but which haven't been linked to any parent genes. After selecting a desirable orphan, Adam scoured the database for similar enzymes in other organisms, along with the corresponding genes. Using this information, it hypothesized that similar genes in the yeast genome may code for the orphan enzyme.

The process might sound simple — and indeed, similar "scientific discovery" algorithms already exist — but Adam was only getting started. Still chugging along on its own, it designed experiments to test its hypotheses, and performed them using a fully automated array of centrifuges, incubators, pipettes, and growth analyzers.

After analyzing the data and running follow-up experiments — it can design and initiate over a thousand new experiments each day — Adam had uncovered three genes that together coded for an orphan enzyme. King's group confirmed the novel findings by hand.

Waltz thinks Adam will inspire other scientists. "They'll realize they can automate more of the process than they currently have. They can explore a wider range of possibilities without doing it all by hand."

King is already expanding his Robot Scientist fleet by producing Eve, which will autonomously design and screen drugs against malaria and schistosomiasis.

"Most drug discovery is already automated," says King, "but there's no intelligence — just brute force." King says Eve will use artificial intelligence to select which compounds to run, rather than just following a list.

If robotic scientists made their way into other labs, their human counterparts would not be out of a job anytime soon. If anything, they may find their work more exciting.

"There may be teams of humans and machines," says King. "Robots will be doing more and more of actual experimental work and simple cycles of hypothesis generation. Humans would migrate to more strategic and creative positions. How can we waste trained post-docs by making them pipette things in labs? It's crazy."

But with advances in artificial intelligence, it's conceivable that the role of robots would, in the more distant future, creep deeper into the human realm, progressing from lab technician to lab head. Robots may even be capable of performing supposed acts of genius, such as Einstein's conception of special relativity.

"There isn't any intrinsic reason why that wouldn't happen," says King. "I think there's a continuum between the really basic types of science that you'd get from Adam, and the things I can do, and then Einstein-type science. A computer can make beautiful chess moves, but it's not doing anything special. It's just doing more of the same thing. In my view that's what's going to happen in science."

King may already have a head start: Deep Blue could never have beaten Garry Kasparov without engineer Feng-Hsiung Hsu moving the pieces on its behalf.

See Also:

Image: Jen Rowland