Linking patent data and scientific research: the innovation accelerator we've all been looking for
12 Nov 2020 10:32 am by Megan Burnside
George Tsatsaronis is Vice-President Data Science Research Content Operations at Elsevier. In his vision, patent data and scientific research are two sides to the same medal. If these two disciplines are combined in an intelligent way, the industry will be able to convert knowledge into marketable products faster than ever before.
At Elsevier, an international publisher of scientific publications, George Tsatsaronis heads up the Data Science and Research Content Operations department. With a background in natural language processing and data mining, he knows exactly how to convert massive amounts of data into transparent, useable information. He made it his life’s work: ‘It all comes down to defining interlinkages and correlations.’
Basis for governance policy and business strategy
There is a clear parallel between Tsatsaronis’ work and the patent solutions from LexisNexis - which all focus on patent data analysis. Tsatsaronis uses artificial intelligence, machine learning and natural language processing to analyse an enormous amount of scientific publications. To this end, he examines the funders plus the amount of subsidies and compares the intended outcomes to the actual outcomes.
The objective is to index the raw data at a large scale, so these can be searchable for both subsidy applicants and subsidy providers. This is important according to
Tsatsaronis, because subsidy providers require full transparency into the funds they make available. Moreover, subsidy applicants like universities want to know which research areas are still underexposed. Tsatsaronis: ‘With the results one can justify the need for subsidies, but the data also forms the basis for government to formulate policies and for businesses to make better informed strategic decisions.’
The necessity of patent data
This is where patent data comes in, says Tsatsaronis. The past years, there has been a shift in the way companies operate. Their competitive advantage is defined by innovation in more ways than ever before. In order to take the right strategic decisions and minimize unnecessary use of resources, we need to have a complete and comprehensive picture of the world surrounding us. That need translates directly into a the necessity to include patentinformation into the data set that is used as the basis to define policy. ‘Where patent data used to be a nice-to-have in the strategy developing process, it now turned into a need-to-have.’
In Tsatsaronis’ view, the time is now to link scientific, peer reviewed research to patent data. Ideas to do this first came to the table in the nineteen-eighties. Still a theory by then, the stormy rise of computer technology - and specifically artificial intelligence - we have seen in the past decades makes this (with the help of superfast computers) practically feasible today. ‘Currently, we are still in the rebirth of the neural networks. That means that we are now in the ideal position to make optimal use of these.’
Linked data is the ‘new gold’
When it comes to transforming these ambitions into reality, Tsatsaronis observes a shift in perspective. Before, when people spoke about ‘the new gold’, they were referring to data in general. Whereas today, linked data is considered to be the new gold, according to Tsatsaronis. Nowadays, collecting data is fairly easy and inexpensive. While that is a great development, he is convinced that the real value is to be found in identifying valuable interlinkages and correlations.
To make the interlinkages in massive datasets with different origins workable, it is essential that you make the right decisions in the strategy development process. Assessing sources and applying hierarchy; it entails a lot of manual work according to Tsatsaronis. And that is not the only problem: a big amount of information is of questionable origin. The crux of this challenge is assessing what is valuable and useful, says Tsatsaronis: ‘That is most important part, but the hardest job at the same time’
Creating opportunities through short cuts
According to Tsatsaronis, we live in a time of change. Before, inventions were given away and research was conducted in collaboration with the industry.
Nowadays, universities tend to issue a patent for an invention themselves. That results in a switch: new inventions often make a first appearance as a patent
and are published as a peer reviewed article in a later stage
How can businesses use this to their advantage? Where are the opportunities when scientific research is linked to patented ideas? In short, it all comes down to creating short cuts, according to Tsatsaronis.
The ideal image: accelerating innovation
As an example, he takes a farmaceutical company that is looking for treatment of a specific form of cancer. By extracting data from a specific genome or from a treatment that has already proven to be successful, researchers can determine the direction with the best chances of a working medicine much faster. 'By doing this, the time-to-market is shortened significantly. That offers enormous opportunities to make new inventions available in a faster way; and this is something that benefits society as a whole.’
Another aspect that makes the connection of scientific data and patent data beneficial to businesses is that a lot of knowledge is gathering dust on
the shelves or buried under piles of research. Providing transparency into all those inventions and combining them with patents, opens a goldmine of opportunities according to Tsatsaronis. However, he stipulates that it is important to make a side note here: this is the ideal image, the future as he envisions it, in which knowledge can be extracted and applied in practice.
Hybrid intelligence – AI collaborating with people
What needs to be done to get at that point as soon as possible? He doesn’t want to speculate. He does notice the availability of better hard and software and rapid decline in the number of obstacles. Developments happen in a faster pace than ever before and are exceeding everyones expectations, says Tsatsaronis. Artificial intelligence even has a successor just around the corner; everybody is already talking about hybrid intelligence. Tsatsaronis: ‘Hybrid intelligence is artificial intelligence that collaborates with humans. It will help us to achieve our goals in a way that is better, easier, faster and better informed.’
Beneficial to society as a whole
Matthijs van der Pol/Jorda Zuurendonk