Search

We review the main components of autonomous scientific discovery, and how they lead to the concept of a Robot Scientist. This is a system which uses techniques from artificial intelligence to automate all aspects of the scientific discovery process: it generates hypotheses from a computer model of the domain, designs experiments to test these hypotheses, runs the physical experiments using robotic systems, analyses and interprets the resulting data, and repeats the cycle. We describe our two prototype Robot Scientists: Adam and Eve. Adam has recently proven the potential of such systems by identifying twelve genes responsible for catalysing specific reactions in the metabolic pathways of the yeast Saccharomyces cerevisiae. This work has been formally recorded in great detail using logic. We argue that the reporting of science needs to become fully formalised and that Robot Scientists can help achieve this. This will make scientific information more reproducible and reusable, and promote the integration of computers in scientific reasoning. We believe the greater automation of both the physical and intellectual aspects of scientific investigations to be essential to the future of science. Greater automation improves the accuracy and reliability of experiments, increases the pace of discovery and, in common with conventional laboratory automation, removes tedious and repetitive tasks from the human scientist.

In their 19 June letter (“Machines fall short of revolutionary science,” p. [1515][1]), P. W. Anderson and E. Abrahams, commenting on our work on the automation of science, state that we are “seriously mistaken about the nature of the scientific enterprise.” Their argument seems to be based on two premises: (i) There are two types of science, normal and revolutionary, and normal science “does not contribute very much to the advancement of knowledge.” This view dismisses as unimportant the vast bulk of science, and must surely be wrong. (ii) Whereas normal science may be automated, revolutionary science never will be, as there is no possible “mechanism.” It is certainly true that revolutionary science cannot currently be automated, and in our Report (“The automation of science,” 3 April, p. [85][2]) we described the automatically generated science as “modest…but not trivial.” Nevertheless, the inability of some critics to imagine a mechanism does not eliminate the possibility that one exists. Indeed, the mechanism we propose is the one that has been successfully applied to chess: There is a continuum in player skill, and computers slowly improved with advances in computer hardware and software until they now play at world championship level. We argue that there is a similar continuum in the ability to do science, from what robot scientists can do today, through what most human scientists can achieve, up to the level of a Darwin or Newton. The Physics Nobel Laureate Frank Wilczek has said that the best chess player in the world is “non-human” and that this may well be true for the best physicist in 100 years time ([ 1 ][3]). Finally, Anderson and Abrahams ignore the possibility of machines and humans working together to do revolutionary science that neither could do alone. 1. [↵][4] 1. F. Wilczek , Fantastic Realities: 49 Mind Journeys and a Trip to Stockholm (World Scientific Publishing, Singapore, 2006), p. 304. [1]: /lookup/doi/10.1126/science.324_1515c [2]: http://www.sciencemag.org/cgi/content/full/324/5923/85 [3]: #ref-1 [4]: #xref-ref-1-1 "View reference 1 in text"

Despite science's great intellectual prestige, developing robot scientists will probably be simpler than developing general AI systems because there is no essential need to take into account the social milieu.