Stay informed with free updates
Just sign up for artificial intelligence myFT Digest — delivered straight to your inbox.
The author is Chairman of Deeptech Labs and Founder of Cantab Capital Partners.
Now, the fairy dust of artificial intelligence is being thrown around. But all the fuss, hype and unprofitable business models may be the limit. Although the word “mathematics” rarely appears in this, we ultimately have to face numbers.
Governments, industry, and markets have largely overlooked the fundamentals of mathematics amidst the frankly insane amounts of money flowing into these technologies. It is past time for the mathematics community to be recognized as a key element in deriving real value and benefit from this field.
Mathematicians often say, “At the end of the day, it’s all about math,” but this is very troubling to experts in fields as diverse as physics, genomics, and economics. But in the case of AI, the relationship is clear.
Modern neural networks effectively work by multiplying tensors (a more general version of matrices and vectors that many people are familiar with) over and over again. Feature extraction in machine learning is done using the mechanism of eigenvectors and eigenvalues. Creating models from data quickly and cheaply is a constrained optimization problem that assumes advanced calculus. And so on.
In fact, AI and its more humble form, machine learning, are subfields of statistics, and statistics itself is a subfield of mathematics. Data is the fuel for AI, mathematics defines the rules for analyzing data, and computer scientists use hardware and software to implement these rules.
But the UK government has failed to understand how important the mathematics community is to our goals of AI excellence, and to extracting real value from it. Since coming to power, the government has cut funding for a number of mathematics initiatives, from the Advanced Mathematics Support Programme, which encourages high performance in schools, to the Multiplication Scheme, which aims to increase mathematical literacy among the general public.
They have also stood by while many university mathematics departments have suffered layoffs or closures. Research from the Mathematics and Science Campaign shows that the proportion of students studying mathematics is decreasing and the number of mathematics graduates is predicted to decline over the next decade, at a time when we need it most.
As we saw in our 2018 Bond Review, this sector offers great bang for your buck. Mathematics is estimated to generate gross value added of around £500bn. Yes, around 20% of UK GDP.
Although some may quibble about its magnitude, it is clear that mathematics contributes significantly to our economy. And while mathematician caricatures, which require only paper and pencil, are a bit outdated, they are still by far the cheapest division of the Stem quad.
Once something is proven, it can be used by other researchers forever. Case in point: an 18th century statistician named Thomas Bayes discovered a theorem that is still the basis for how many AI systems “learn” from new data.
Britain has traditionally far outperformed its own peers in mathematics. However, if we don’t continue and expand our support for this subject, we can build as many data centers as we want, but the innovation that drives what happens there will come from elsewhere. Mathematics is the most cost-effective route to achieving intellectual, ethical, and technical leadership in the field.
Even if the AI bubble bursts, value will still be extracted from the rubble. Applying mathematics to real-world problems in medicine, government, business, and our personal lives will continue to pay dividends, just as it does in the world of finance. Even in a catastrophic scenario, mathematics will continue to be the engine of future AI innovation, so we need to support it.
This article was amended after its original publication to correct the era in which Thomas Bayes lived.
