MIT researchers have put AI to great use. The team has predicted which technologies have the potential to grow and which ones are just hyped out of proportion. In a new study, they quantitatively assessed the future potential of 97% of the US patent system and found that the fastest-improving domains were mostly software-related.
Their findings were converted into an online system where users can feed in keywords to know improvements forecasts for specific technologies. Their research could allow entrepreneurs, investors, policy-makers, and researchers to understand the future of opportunities in the technology world.
New algorithm used for the study
“Our method provides predictions of performance improvement rates for nearly all definable technologies for the first time,” said lead author Anuraag Singh in a statement.
For the study, the team used a new probability-based algorithm along with machine learning, and patent network analytics to predict how different technologies would improve and at what rate. The researchers first divided the patents into 1,757 discrete technology domains. Each of them had inventions that fulfill a specific function with the aid of different branches of scientific expertise. The team then estimated the average “centrality” of patents for all domains.
“Central patents are like information hubs in the citation network, representing inventions that are related technologically by a path of improvements to many other inventions that appeared before and after them,” reads the study paper.
Good improvements were forecast
The rate of improvement differed from 2% per year for “Mechanical skin treatment — Hair removal and wrinkles” to 216% per annum for “Dynamic information exchange and support systems integrating multiple channels.” Averaged out, tech improvements were forecast at a rate of 19% per year.
“The domains that show improvement rates greater than the predicted rate for integrated chips, from Moore’s law, are predominantly based upon software and algorithms,” the researchers wrote. “In addition, the rates of improvement were not a strong function of the patent set size.”
