Microsoft has developed a machine reading comprehension technology that allows computers to read all the information on a given subject and summarize it back within a few minutes. The technology uses artificial intelligence (AI) to read a document and answer the questions about it, as well as a human, according to the company.
In my view, it's not so much about reading faster than human, but the ability to read, comprehend and process the data and then spit it out in a search query. The goal would be to compile all the information and summarize it in a matter of seconds, similar to the way a search engine returns a query of information from the sites it crawled.
Microsoft says the technology will enable its search engine Bing and intelligent assistant Cortana to interact with people and provide information in more natural ways, similar to the way that humans communicate.
Using the Stanford Question Answering Dataset (SQuAD), a team at Microsoft Research Asia in early January submitted a model that reached the score of 82.650 on the exact match portion of the test. The human performance on the same set of questions and answers came in at 82.304.
On January 5, researchers with the Chinese e-commerce company Alibaba submitted a score of 82.440, also about the same as a human. Now Microsoft and Alibaba are tied for first place on the SQuAD “leaderboard,” which lists the results of research efforts.
It's the start of being able to create more technology that people can interact with in simple and intuitive ways, according to Microsoft. For example, per Microsoft, instead of typing in a search query and getting a list of links, Microsoft’s Bing search engine is moving toward efforts to provide people with more plain-spoken answers, or with multiple sources of information on a topic that is more complex or controversial.
With machine-reading comprehension, researchers say computers also would be able to quickly parse through information found in books and documents and provide people with the information they need most in an easily understandable way.
The biggest challenge I see is making sure the information is current and not outdates. So how can Microsoft, or any company using this type of technology, recognize when the information is outdated if it is found in books to serve up the most accurate information?