In 1986, Don R Swanson argued that there is a lot of undiscovered public knowledge, i.e. publicly available complementary but *not interactive* academic literatures /1 https://www.jstor.org/stable/4307965?seq=1

the growth of science has led scientists to organise their work into specialities. Specialities that grow too large tend to divide into subspecialties that have their own literature /2

This has led to an incalculably greater combinatorial explosion of *unnoticed* and unintended logical connections between academic pieces of literature. /3

For example, suppose one (biomedical) literature
establishes that some environmental factor A influences
certain physiological conditions and a second
literature establishes that these same physiological
changes influence the course of disease C. /4

Presumably, then, anyone who reads both literatures could conclude that factor A might influence disease C. If, however, the two literatures were developed independently of one another, the logical linkage illustrated maybe both unintended and unnoticed /5

it is possible that scientists have not previously considered
both literatures together, and so it is possible that no one is
aware of the implicit A-C connection. How do we fix this? /6

Swanson has developed principles and tools for tackling this problem. He also wrote a 10-year update of the initial paper where he mentions several examples /7 https://aaai.org/Papers/KDD/1996/KDD96-051.pdf

How does a 25-year update look like and how will advances in entity linking and semantic search help realise Swanson's vision and give us a new set of hypothesis to test? /8

For this to work you need to search the whole literature space with good amount of compute. Why isn't this a module within @SemanticScholar or @MSFTAcademic ? Why is the scientometric folk working on posthoc analysis rather than generative solutions in hypothesis generation?

@logangraham