Subscribe now

Mathematics

DeepMind created a maths AI that can add up to 6 but gets 7 wrong

By Adam Vaughan

4 April 2019

Maths exam

No calculators allowed

Thomas LENNE/Alamy

Artificial intelligence firm DeepMind has tackled games like Go and Starcraft, but now it is turning its attention to more sober affairs: how to solve school-level maths problems.

Researchers at the company tasked an AI with teaching itself to solve arithmetic, algebra and probability problems, among others. It didn’t do a very good job: when the neural network was tested on a maths exam taken by 16-year-olds in the UK, it got just 14 out of 40 questions correct, or the equivalent of an E grade.

There were …

To continue reading, subscribe today with our introductory offers

View introductory offers

No commitment, cancel anytime*

Offer ends 14th June 2023.

*Cancel anytime within 14 days of payment to receive a refund on unserved issues.

Inclusive of applicable taxes (VAT)

or

Existing subscribers

Sign in to your account

More from New Scientist

Explore the latest news, articles and features

Popular articles

Trending New Scientist articles

1

Which countries will do well at Eurovision 2023 – according to science

2

Chaotically shifting planets could be a sign of advanced aliens

3

Electromagnetic fields from power lines are messing with honeybees

4

Emergence: The mysterious concept that holds the key to consciousness

5

A distant supernova defies our understanding of the cosmos’s expansion

6

The mental tricks you can use in your lifelong pursuit of happiness

7

We now know how parasitic worms help ward off obesity and diabetes

8

Why the causes of poor mental health may share a common root

9

Why the origins of life remain a mystery – and how we will crack it

10

Lab-grown meat could be 25 times worse for the climate than beef