Imagine an AI that doesn't just solve math problems, but helps invent new math. That's the wild idea behind a new free tool called Axplorer. It's designed to help mathematicians spot patterns nobody's ever seen before, right from their Mac Pro.
This isn't just about crunching numbers faster. It's about breaking new ground in math. Axplorer is a souped-up version of an older tool, PatternBoost, which famously cracked a tough math puzzle called the Turán four-cycles problem. That original tool needed a supercomputer. Now, anyone can download Axplorer and get that same power.
The AI That Thinks Differently
Most AI tools are great at finding answers to problems that already exist. Think of them as super-smart students who ace every test because they've studied all the answers. But math isn't always about known answers; it's about exploration.
We're a new kind of news feed.
Regular news is designed to drain you. We're a non-profit built to restore you. Every story we publish is scored for impact, progress, and hope.
Start Your News DetoxFrançois Charton, a research scientist at Axiom Math, explains that other AI models, like the big language models (LLMs) you hear about, are good at tasks similar to things they've already seen. They're "conservative," he says, because they tend to reuse existing ideas.
But real breakthroughs in math often need completely fresh ideas. They come from spotting patterns no human has ever considered. That's where Axplorer shines. You give it an example, it generates similar ones, you pick the interesting ones, and it learns to find even more. It's like having a brainstorming partner who can explore millions of ideas in minutes.
Solving "The Big Problems"
Charton isn't interested in the easy wins. He wants to tackle "the big problems that have been very, very well studied and famous people have worked on them." Last year, Axiom Math used another tool, AxiomProver, to solve four such difficult math problems. The Turán four-cycles problem, for instance, is a big deal in graph theory – the math behind understanding complex networks like social media or supply chains.
When PatternBoost first solved the Turán problem, it took three weeks on thousands of supercomputer machines. Charton called it "embarrassing brute force." Axplorer, however, matched that result in just 2.5 hours, running on a single machine. That's seriously cool.
Carina Hong, Axiom Math's founder, says their goal is to make these powerful tools accessible. The code for Axplorer is open source, meaning anyone can use and build on it. She hopes students and researchers will use it to speed up mathematical discovery, helping us find the next big thing in technology and beyond.
