More science doesn't always mean better science. Fashionable theories, pressures to publish and cultural biases can lead to reduced innovation even as the volume of journals publishing scientific findings grows exponentially. With government austerity creating tighter science budgets and increased demand for accountability, it's more important than ever that we find ways of evaluating and forecasting the potential of a scientific finding to add important information to the body of knowledge, argues CI fellow James Evans in the journal Science.
Last month, Evans, the director of the CI's Knowledge Lab, speculated on these topics as part of a commentary on a new model that predicts the citations and scientific impact of publications. The model, developed by authors from Northeastern University, adapting measures that search engines use to assess website popularity and evolutionary biology terms ("fitness," or the importance of a paper relative to its peers) to predict an article's impact over time. Evans argues that such models could change the practice of science as we know it, throwing dollars and expertise behind ideas with a higher potential for success and bringing laboratory results into real world applications more quickly.
The ability to better predict an article's success could translate into a faster scientific life cycle for the discovery—from time of publication to widespread acceptance. This might then translate into faster convergence to best practices that would boost the number of scientists with skills required to build on an impactful discovery.
Evans also envisions a future where models can be applied to predict not just the future of publications, but also the future of scientists and ideas. Improved methods of mining publications for their results and claims (such as Peter Murray-Rust's Content Mine project to "liberate 100 million facts from the scientific literature") may also lead to "robot scientists" who come up with the most fruitful hypotheses.
In this way, citation prediction represents one step on the path to creating algorithmic or robot “scientists” that are more creative, risky, persistent, and wide-reading than ourselves. By enabling scientists to consider not only the most fruitful hypothesis but also the most fruitful algorithm for generating hypotheses, future prediction methods would augment scientific ability, increase productivity, and multiply returns from science for society.
For more on Evans' work studying the patterns of knowledge and science, visit the Knowledge Lab website.
[Image: "Robot at the Museum of Science and Technology" by Peter Lindberg, from Wikimedia Commons]