Robot bookworms? Artificial intelligence (AI) can now read as well as humans after Microsoft and Alibaba made research breakthroughs
The march of artificial intelligence continues…
Researchers at Microsoft have hailed a “major milestone” for the technology after they managed to create a system using the technology that can read just about as well as a human.
They used what’s known as the Stanford Questions Answering Dataset (SQuAD), a tool commonly employed in scientific research to assess machine reading comprehension, achieving the best ever score of 82.304, on a par with human comprehension.
Read more: Robots could lead to greater wage inequality, think tank warns
In addition to reading the information, the technology can also be quizzed on what it’s about, answering questions, with implications for search engine Bing and voice assistant Cortana. The tech company is already looking at applying the technology to those products.
And Microsoft isn’t alone in the breakthrough. Chinese tech giant Alibaba also recorded a similar score of 82.440, putting the two tech giants jointly at the top of the SQuAD leaderboard.
According to Microsoft, the technology could be used to make books and documents digital – essentially making the “control F” function on computers for finding words and phrases applicable to printed information.
They said it would would, for example, “let drivers more easily find the answer they need in a dense car manual, saving time and effort in tense or difficult situations”.
“These tools also could let doctors, lawyers and other experts more quickly get through the drudgery of things like reading through large documents for specific medical findings or rarified legal precedent. The technology would augment their work and leave them with more time to apply the knowledge to focus on treating patients or formulating legal opinions,” they said.
Read more: This bank is using artificial intelligence to help, not replace traders
Microsoft Research Asia assistant managing director Ming Zhou said the milestone was “just a start” however. “Natural language processing is still an area with lots of challenges that we all need to keep investing in and pushing forward,” he said.
And Alibaba’s chief scientist for natural language programming Luo Si said the technology “can be gradually applied to numerous applications such as customer service, museum tutorials and online responses to medical inquiries from patients, decreasing the need for human input in an unprecedented way.”