AI: Beating humans in Math

...by crossing the probabilistic and deterministic calculation streams

For the longest time, even years before the year plus old ChatGPT moment driven by the Foundation LLM AI model GPT by OpenAI, we’ve been conditioned of never-ending tales of computers getting exponentially smart enough to beat humans. From Chess by IBM’s Watson, to the Asian game Go by Google AlphaGo AI teams, to many more.

And of course since the advent of LLM AI and Generative AI models now sparking the AI Tech Wave and related AI infrastructure ‘gold rush’, barely a week goes by when we don’t hear of AI passing high level law exams, coding tests, accounting tests and so many other skills that for now have been deemed only good enough for human intelligence.

So at first glance, a NYTimes headline “AI’s Latest Challenge: the Math Olympics”, just seems to be another in a similar long line of computers beating humans at ‘human’ intelligence determining skills. The piece even comes with its own headline warning:

“Watch out, nerdy high schoolers, AlphaGeometry is coming for your mathematical lunch.”

Followed by the context:

“For four years, the computer scientist Trieu Trinh has been consumed with something of a meta-math problem: how to build an A.I. model that solves geometry problems from the International Mathematical Olympiad, the annual competition for the world’s most mathematically attuned high-school students.

Last week Dr. Trinh successfully defended his doctoral dissertation on this topic at New York University; this week, he described the result of his labors in the journal Nature. Named AlphaGeometry, the system solves Olympiad geometry problems at nearly the level of a human gold medalist.”

The piece goes onto to highlight the benchmarks that AlphaGeometry achieved vs humans:

“Dr. Trinh presented the AlphaGeometry system with a test set of 30 Olympiad geometry problems drawn from 2000 to 2022. The system solved 25; historically, over that same period, the average human gold medalist solved 25.9. Dr. Trinh also gave the problems to a system developed in the 1970s that was known to be the strongest geometry theorem prover; it solved 10.”

But here’s where this story goes beyond just ‘plain old’ LLM AI based on ‘probabilistic computing, a topic I’ve written about at length:

“AlphaGeometry is a “neuro-symbolic” system. It pairs a neural net language model (good at artificial intuition, like ChatGPT but smaller) with a symbolic engine (good at artificial reasoning, like a logical calculator, of sorts).”

This is a direction we’re likely to see being taken going forward, something I’ve.described in a post last November “Doing the Reasoning Math Reliably, crossing the probabilistic and deterministic streams”. Specifically:

“In a piece titled ”OpenAI made an AI breakthrough before Altman firing, stoking excitement and concern”, the Information provides color on an OpenAI research development that addresses the holy grail issue of AI software vs traditional software in these early days of the AI Tech Wave. It potentially crosses (fusing) the probabilistic (aka stochastic) approach of the former with the deterministic approach of the latter. “

“The current holy grail of LLM AI Research is fusing the uniquely probabilistic way AI produces its results vs the deterministic way computers have produced reliable answers doing math with 0s and 1s. In a piece “AI: Can’t see how it works” back in June, I underlined”:

“One of the hardest things for regular folks to get their head around is how AI software is different from traditional software. Why is AI eating software?”

“The way I’ve tried to explain AI software before is to highlight that traditional software is based on deterministic code vs AI software being run as probabilistic code. Of course on super fast parallel chips (GPUs), using massive amounts of reinforcement learning feedback loops, constantly feeding on all our Data ever put online, to make the eventual results of the Foundation LLM AI models more reliable and relevant. Less hallucinatory.”

“And how all this makes this AI Tech wave far more potent in human opportunities than any other tech wave before it like the PC or the Internet.”

“Computer software and hardware traditionally does rock solid math. Utterly reliable to the umpteenth decimal place every single time.”

The way humans have done math for centuries.“

I highlight all this because the AlphaGeometry development continues down this path of ‘blending the streams”.

Between probabilistic and deterministic math, all run by ever more powerful computers. These increasingly impressive capabilities by AI and computers are but an early step towards this fusion of two different ways of doing math using software and ever more powerful AI hardware.

We should be less focused om how AI is ‘beating’ humans young and old. And more focused on how these technologies will make all humans a lot better doing more things with far harder and tougher math. Just like we did with calculators. Stay tuned.

(NOTE: The discussions here are for information purposes only, and not meant as investment advice at any time. Thanks for joining us here)