AI is at its best when it has a huge amount of data to sort through, and uses smart algorithms and tremendous data crunching capability to make sense of that data for us.  From the original Google search engine to the AI diagnosis tools that are entering healthcare at the moment, that’s the basic proposition.

With thousands of academic papers published every day, together with fascinating new startups springing up around the world, the task of any innovation or futurism department is an increasingly challenging one.  A number of tools have entered the market to try and help, and I’ve covered a few of them.

For instance, Meta is a tool that emerged out of SRI International.  The service trawls the academic literature from over 17 million academics and uses things like trajectory mapping to underpin its AI engine and provide users with what it believes are the hot trends in any given field.

Or you’ve got Konfer, which was developed by a UK based consortium to do a similar job, albeit pretty badly in my opinion.


On a recent trip to the Grenoble Ecole De Management I learned about another research support platform, called Expernova.  The Montpellier based company have been going for around ten years, but it’s perhaps fair to say that the convergence of both available data and the computing power to make sense of the data has supercharged their capabilities.

Their database has data from 52 countries around the world, including research articles, patents, clinical trials, whitepapers, grants and projects from the likes of Horizon2020 and a burgeoning database of startup information.

As a tool, it’s incredibly comprehensive, and in a short test-run, it returned the kind of results you would hope for.  As a French company, many of their clients are French multinationals, with the likes of Nestle and L’Oreal on their roster, but the company are keen to expand internationally.

Known unknowns

Will they transform the way innovation is performed?  Yes and no.  In terms of the research they allow you to undertake, it’s fantastic, and they provide accurate insight into who both the best researchers are, the best institutions and the best startups.

My main reservation is that you need to actively know what it is that you’re looking for.  You need to know that you want to innovate with graphene or Crispr.  It doesn’t really take into account the recombinations that are increasingly important for innovation.  Or in other words, the innovations that already appear in one field and you apply elsewhere.  The unknown unknowns if you like.

A paper from a few years ago looked at the importance of this form of innovation.  The paper reveals that throughout the history of the USPTO, roughly 40 percent of all patents are actually refinements of existing work, with the remainder therefore being novel works.  The trend has moved inexorably towards recombination, with the vast majority of current innovations building on pre-existing work.

Suffice to say, that isn’t to say that AI won’t eventually automate that recombinative function as well.  For instance, a recent paper from researchers at Carnegie Mellon University and the Hebrew University of Jerusalem highlights an AI driven approach to mine databases of patents and research papers for ideas that can be recombined into solutions for new problems.

Central to the approach was an attempt to find analogies that connect seemingly disparate methods and problems. They used crowdsourcing to understand how people form analogies, before using this to train a deep learning algorithm to then mine the intellectual databases for potential innovations.

This functionality isn’t built into tools such as Expernova yet, but it wouldn’t surprise me to see support for this in the not too distant future.