The researchers at MIT have built an artificial intelligence (AI) framework to be able to share an 'early-alert' signal to recognise future high-impact technologies, by observing patterns from previous scientific publications. 

DELPHI, short for Dynamic Early-warning by Learning to Predict High Impact, has already been tested to ensure that it can in fact identify potentially pioneering papers. It was in fact able to identify all the upcoming papers on an experts’ list of key seminal biotechnologies, sometimes as early as the first year after their publication.

The Media Lab’s Molecular Machines's research group, James W. Weis, a research affiliate of the MIT Media Lab, and Joseph Jacobson, a professor of media arts and sciences and head of the research group, tested 50 recent scientific papers to predict which ones would have high impact by 2023. The papers covered subjects such as DNA nanorobots used for cancer treatment, high-energy density lithium-oxygen batteries, and chemical synthesis using deep neural networks, among others. 

Researchers think that DELPHI can help researchers prioritise funding for potentially effective technologies that would otherwise have been hidden in the mountain of discoveries. The researchers also think that governments, philanthropies and venture capital firms stand to gain infinitely as they would know which scientific advancements to support. 

“In essence, our algorithm functions by learning patterns from the history of science, and then pattern-matching on new publications to find early signals of high impact,” says Weis in a blog released on MIT's website. “By tracking the early spread of ideas, we can predict how likely they are to go viral or spread to the broader academic community in a meaningful way.” The paper has been published in Nature Biotechnology.

The machine learning algorithm scans all the vast digital information related to scientific publications in scientific publications since the 1980s, drawn from 42 biotechnology-related journals, including over 7.8 million individual nodes, 201 million relationships and 3.8 billion calculated metrics. 

DELPHI has demonstrated the ability to chart the papers, authors, institutions and other such data to create a knowledge graph that can build connections to determine their properties. “These nodes and edges define a time-based graph that DELPHI uses to learn patterns that are predictive of high future impact,” explains Weis.

Together, these network features are used to predict scientific impact, with papers that fall in the top 5 percent of time-scaled node centrality five years after publication considered the “highly impactful” target set that DELPHI aims to identify. 

The DELPHI can be used in “incentivizing teams of people to work together, even if they don’t already know each other — perhaps by directing funding toward them to come together to work on important multidisciplinary problems,” said Weis.

"Advancing fundamental research is about taking lots of shots on goal and then being able to quickly double down on the best of those ideas,” says Jacobson. “This study was about seeing whether we could do that process in a more scaled way, by using the scientific community as a whole, as embedded in the academic graph, as well as being more inclusive in identifying high-impact research directions."

The researchers revealed that DELPHI was able to recognise exceptional papers within a year of publishing that would go on to have a significant impact later. However Weis has a warning to not consider DELPHI as a fortune teller predicting the future. “We’re using machine learning to extract and quantify signals that are hidden in the dimensionality and dynamics of the data that already exist.”

Hopefully, DELPHI will be less biased towards evaluating research as it looks beyond just the citations and journal factor numbers, which can be manipulated; it also considers a paper's appeal on a wider, more diverse audience. However, once again Weis has a warning. “We need to constantly be aware of potential biases in our data and models. We want DELPHI to help find the best research in a less-biased way — so we need to be careful our models are not learning to predict future impact solely on the basis of sub-optimal metrics like h-Index, author citation count, or institutional affiliation.”

The scientific diaspora is very excited about this breakthrough because they believe DELPHI can guide scientific investing to become a lucrative avenue to build financial products as identifying groundbreaking work becomes more efficient and effective. 

“The emerging metascience of science funding is pointing toward the need for a portfolio approach to scientific investment,” notes David Lang, executive director of the Experiment Foundation. “Weis and Jacobson have made a significant contribution to that understanding and, more importantly, its implementation with DELPHI.”