Adding one irrelevant sentence to math problems causes AI systems to make confident mistakes over 300 percent more.

GSM8K-V is a purely visual multi-image mathematical reasoning benchmark that systematically maps each GSM8K math word problem into its visual counterpart to enable a clean, within-item comparison ...

January 8, 2026 - It's time for John Fensterwald's annual predictions for what's in store for education in 2026. Math is the sum of its parts, and it adds on itself. What does that mean? It means that ...

Here's the thing about math that nobody tells you: it's less about memorizing formulas and more about knowing which tools to reach for. By fourteen, students should have a problem-solving toolkit that ...

“Yoshua recently turned 57. He is three years younger than Yann. How old is Yann?” Solving such a math word problem (MWP) requires understanding the short natural language narrative describing a state ...

In the third century BCE, Apollonius of Perga asked how many circles one could draw that would touch three given circles at exactly one point each. It would take 1,800 years to prove the answer: eight ...

"Math Homework Hotline" has been solving problems for local students for 33 seasons. Hillsborough County Schools produce the show and help kids in all grades. ‘Americans are going to feel it’: Bessent ...

So far, scientists have relied on positive reinforcement learning to train LLMs, but the opposite seems to be giving much better results, finds Satyen K. Bordoloi… This is a finding that’ll have ...

When the greatest mathematician alive unveils a vision for the next century of research, the math world takes note. That’s exactly what happened in 1900 at the International Congress of Mathematicians ...