Summary of awarded fellows and program outputs
In 2023, we awarded five fellowships to PhD students in the field of innovation research. Of these five, four were also able to join us at the December Technical Working Group meeting, where they presented their research to a broad audience.
The projects covered a diversity of topics and techniques, and included a number of data sources we had not seen commonly used.
Patents or Defensive Disclosures?
Building a BERT classifier to match defensive disclosures with patent technology classes (CPC classes).
Pydrad
Development of an open-source Python package to construct, transform, combine and compare well-used innovation datasets. Supports and streamlines analysis, and also provides a useful index.
Linking Scientific Articles to Media Mentions
Linking EurekaAlert data to author data on OpenAlex, using CrossRef dates to increase accuracy. Using the media landscape of science promotion to understand press impact of innovation and scientific discoveries.
Statistical Evidence from Clinical Trials
Using LLMs, to combine clinical trial records with scientific publication and FDA approvals, creating a new dataset of clinical trial results.
Analysis of Knowledge Entities
Using Bio-BERT as an example to developing a metric for innovation impact based on “knowledge entities”. As opposed to citation, this method attempts to understand the change in ‘attention’ directed at one such entity as an impact measure.