There’s been much note in the blogosphere in recent days about Aberystwyth University’s Robot Scientist project and its Adam robot:

Since the project has been going on for quite a few years, the reason it’s suddenly in the news is presumably the publication of a report, The Automation of Science, in Science.

The basis of science is the hypothetico-deductive method and the recording of experiments in sufficient detail to enable reproducibility. We report the development of Robot Scientist “Adam,” which advances the automation of both. Adam has autonomously generated functional genomics hypotheses about the yeast Saccharomyces cerevisiae and experimentally tested these hypotheses by using laboratory automation. We have confirmed Adam’s conclusions through manual experiments. To describe Adam’s research, we have developed an ontology and logical language. The resulting formalization involves over 10,000 different research units in a nested treelike structure, 10 levels deep, that relates the 6.6 million biomass measurements to their logical description. This formalization describes how a machine contributed to scientific knowledge.

King et al preprint (PDF)

To my mind, while the construction of the physical machine is quite impressive and will save some graduate students some drudgery, the really important aspect of the work is that they had to formalize the ontology for enough of the field — “the discovery of genes encoding orphan enzymes in S. cerevisiae: enzymes catalysing biochemical reactions thought to occur in yeast, but for which the gene(s) encoding them have not been identified” — for the machine to formulate and test hypotheses about it. Science has always been about formalizing our knowledge of the world. Adam is a small step in the upcoming giant leap of mechanizing that formalization, after which science as a whole will be transformed.

Formalizing knowledge has been one of the things AI has tried to do, most famously in Cyc, and there are reasons to think that this is not the way, or at least the entire way, to a working AI. But there are a lot of things involving common sense that can be left out of a science knowledge base. Another new project that apparently is aimed at such a knowledge base is Wolfram|Alpha, of which the best description I’ve seen in in this interview. One of these may evolve into the framework into which the formal ontologies and knowledgebases from projects like Robot Scientist ultimately fit.