mergeflow provides deep insights into the next generation of technology by creating a unique knowledge graph. Large global tech corporations and investors use this graph to search for specific technologies and discover the products, research, patents, start-ups, and experts associated with it.
mergeflow builds this graph with information from a broad cross-section of sources that have been fed through its proprietary natural language and machine learning algorithms to determine trends and associations. Using Mergeflow’s knowledge base, global tech corporations can innovate future products and technology, and investors can lead the pack in finding and investing in new tech.
mergeflow collects this information for its algorithms by creating streamlined and standardized kimono APIs off of hundreds of sites, from university research centers, patent offices, regulatory bodies, press outlets, tech publications, research projects, databases and blogs etc. Using kimono, Mergeflow can collect information from a much broader set of sources and focus their efforts on the natural language and machine learning algorithms that provide insight and the graph that maps the next generation of technology.
Read more about mergeflow here. mergeflow wrote a great tutorial blog about how they used kimono to create an API off of “Bayerische Patentallianz”, a Bavarian government organization that hosts interesting technology offers that come out of Bavarian research institutions.
As an example of the kinds of insights you can get when you combine kimono and mergeflow, we looked at the following two questions:
2. Technology offerings: which are prominent research organizations with technology offerings in the area of natural language processing? And at these organizations, who are the relevant people?
In order to address these two questions, we collected information from a broad variety of sources (investor news, university publications, technology licensing offers, etc.). Then, mergeflow algorithms extracted investor names and their associations with Technion and TU Munich, and organization and people names and their associations with “natural language processing”. We then displayed the results as interactive visualizations in mergeflow.
Here is a screenshot of what we got regarding our first question, comparing Technion and TU Munich investor networks:
Note that in order to make a fair comparison, we searched for all possible name variations of TU Munich; “Technical University of Munich”, “TU Munich”, etc.. For Technion, at least to our knowledge, there are no synonyms.
Comparing these two networks, we garnered two preliminary insights:
* Technion’s investor network is clearly bigger than the one of TU Munich.
* Looking at the names of investors associated with Technion suggests that Technion’s network is not only bigger but more internationally distributed.
As a small caveat, we we did not consider Hebrew documents here which may have skewed results against Techion, simply because we used fewer documents for Technion than for TU Munich.
*Technology offerings in natural language processing*
Now let’s turn to our second topic, finding organizations with technology offerings in the area of natural language processing (NLP). We first looked at an interactive timeline in mergeflow, which shows NLP technology offering activities by research organizations over time:
The more technology offerings by an organization, the bigger the green bubble. Columbia (or rather their technology transfer organization, Columbia Technology Ventures) features prominently in our timeline, so we zoomed in on them.
As mentioned above, we were also interested in relevant people. So we displayed an interactive network graph, showing our queries in yellow and people names identified by mergeflow in red:
These are people associated with NLP technology offerings at Columbia (there are other NLP experts at Columbia too, but these are the names associated with technology offerings). If we wanted to pursue this further now and dig deeper, we could, for instance, look at patents, publications, or company activities associated with these people.
About the authors:
Andreas is an analyst and touchpoint for customer relations at mergeflow. After temporarily working in academia and for the German Bundesbank, he decided to join the young and ambitious mergeflow team. He is currently finishing his Master’s Degree in Economics at University of Lausanne (Switzerland) and works part-time from there.
Florian is responsible for company strategy and for analytics development at mergeflow. Before co-founding mergeflow, he worked in software development for operational risk management at institutional investors. He also worked as a Research Associate at the University of Cambridge (Department of Computer Science and Department of Genetics). Florian has a PhD in Cognitive Sciences from MIT.