Connecting Science to Innovation: Open Data and Machine Learning Approaches

Matt Marx
Cornell University, USA, and NBER
 

Abstract:

Prof. Marx will discuss advances in open data, machine learning and their role in innovation research. He will expand on one of his latest projects, which systematically links scientific publications to technological outputs. In this project, the authors provide a dataset of Patent–Paper Pairs (PPPs) across all fields of science, identifying instances where authors of scientific papers also exploit their discoveries in patented inventions. To do so, they train a random forest model based on a combination of hand-checked PPPs. The dataset is then used to revisit the perennial question of whether the patent system fulfills its objective to “promote the progress of science”. Prof. Marx will then conclude by reflecting on how machine learning, and in particular Large Language Models (LLMs), can be leveraged to open new opportunities for large-scale, data-intensive research on science, technology, and innovation.

A discussion will follow by Gianluca Tarrasconi, Chief Data Officer at ipQuants AG, who will offer his perspective on the potential of patent data for innovation analysis and share insights from his startup experience applying LLMs to patent databases to develop specialized reporting and analytics services.

Matt Marx is the Bruce F. Failing, Sr. Chair in Entrepreneurship at the Cornell SC Johnson College of Business, and is the inaugural Faculty Director of Entrepreneurship@Cornell. He leads the Innovation Information Initiative (I3), curates several open datasets at relianceonscience.org, serves as Department Editor for Innovation & Entrepreneurship at Management Science, and is a Research Associate at the National Bureau of Economic Research (NBER). Matt was previously an executive and inventor at two successful startup companies and holds six patents.

Register here.