Scientists at China’s University of Science and Technology of China and the Microsoft Research in Beijing have built a deep learning machine that outperforms an average human on the verbal analogies portion of the IQ test — normally the hardest part of the test for computers.
In particular, they’ve solved the problem of one word having different meanings in different contexts.
Eventually, this kind of technology will lead to better language processing by personal assistants like Siri and better answers from search engines.
Read more about it at the MIT Technology Review.
Latest posts by Maria Korolov (see all)
- Analysts predict drop in headset sales this year - March 25, 2025
- OSgrid enters immediate long-term maintenance - March 5, 2025
- OSgrid wiping its database on March 21: You have five weeks to save your stuff - February 15, 2025