A team led by an Indian-origin researcher at the Massachusetts Institute of Technology (MIT) has developed a new artificial intelligence (AI) tool that can predict how fast any technology is improving, and help organisations manage research and development (R&D).
The study, recently published in the journal Research Policy, demonstrates the predicted improvement rates for 1,757 technology domains covering 97.2 per cent of the US patent system.
The researchers noted that the fastest-improving technologies are predominantly software-related, especially those linked to network management, enterprise security, and media transmission.
"As technologists, we have to choose between technologies all the time. Many times few technologies end up capturing a lot of attention due to the hype around them, said study lead author Anuraag Singh from MIT.
"However, fast-improving, more promising technologies end up being ignored. We want to bring realism and objectivity to technological decision-making," Singh said in statement.
The researchers have also launched a startup called Technext to help organisations anticipate technological advances, breakthroughs, and disruptions.
They also built a free, publicly accessible tool for academic readers to use the system.
The study builds on about two decades of research at MIT, started by Professor Christopher L Magee, one of the co-authors of the study.
It used machine learning, patent network analytics, and empirical technology performance data to predict the future potential of any technology.
The researchers, including Giorgio Triulzi from the University of The Andes in Columbia, note that the key technical innovation in the study is the early identification of technological breakthroughs, bottlenecks, and threats.
This helps organisations avoid being obsolete, improve R&D effectiveness and increase revenue through new product or service development, they said.
The tool can be handy in deploying the USD 2.2 trillion in annual global R&D funding more efficiently by helping organisations avoid low-potential over-hyped technologies and research, the researchers added.
