Throughout history, a small number of exceptional individuals have had a profound impact on our society. However, little is known about how these ‘geniuses’ develop. Here, we propose to build on our Big Data approach which quantitatively defined exceptional performance in science, to create a predictive model of future performance.
We plan to focus on (i) a scientist’s performance, (ii) professional recognition, (iii) societal recognition, and (iv) canonization with the genius label. We will explore multiple data features including bibliometric features, prizes, collaboration, media coverage and others, to identify the temporal patterns predicting exceptional performance. We will further explore multiple biases that impact the emergence of genius, such as the importance of a unique backstory, geographic bias, gender bias, and bias based on publication venues (e.g., top journals). We also plan to incorporate aspects of support, especially that which involves research funding.
A key feature of our approach is that we gather data on both exceptional performers and their peers active in the same creative domain who were not recognized, enabling us to determine unique aspects of exceptional performance.
Ultimately, our goal is to combine our key insights and the training data into a predictive model utilizing recent advances in AI/machine learning to identify individuals already excelling in the areas that predict exceptional performance. We will then ask the developed predictive model how we may be able to foster genius by impacting funding, access to top-tier publication venues, collaborations, media coverage, and other model inputs. Motivated by the Foundation’s longstanding interest in cultivating genius, this research provides a unique opportunity to use a quantitative approach to predict and enhance future performance.
