New AI-based learning system provides personalized math instruction for students

Recognizing how numbers relate to each other is key to mastering math. This ability, known as "structure sense," helps students break down complex problems. Researchers have long studied how early math skills predict future success. But a critical question remains—how do young children develop this ability?

A new tool using eye-tracking technology is shedding light on this mystery. By monitoring how students visually process math problems, scientists can see which strategies they use. This insight could revolutionize how educators identify and support struggling learners.

The idea of structure sense dates back to 1999, when researchers first described it as recognizing patterns within a math problem. Later studies linked early structure sense to arithmetic competence. A first grader with a strong grasp of number relationships is more likely to excel by second grade.

Despite this, little is known about how young students actually display structure sense in their work. Traditional assessments rely on verbal explanations or written responses, but these methods miss unspoken problem-solving strategies. That’s where eye-tracking technology comes in.

Eye tracking follows a person’s gaze to see where and how long they focus on specific parts of a problem. When solving math tasks, children’s eye movements reveal their thought processes. This method captures subconscious strategies that students may not be able to explain.

"Tracking eye movements in a single system using a webcam, recognizing learning strategies via patterns, and offering individual support is completely new," says Maike Schindler, a professor of mathematics education.

Recent research shows that even five-year-olds recognize numerical patterns using visual cues. In one study, children who quickly identified groups of dots solved problems faster than those counting each dot individually.

Another study found that ten-year-olds who focused on entire groups rather than individual numbers performed better in arithmetic. These findings, published in Springer Nature, suggest that students develop a natural sense of mathematical structure over time.

However, researchers wanted to explore structure sense in broader contexts. Instead of just testing how children count objects, they examined how they compare numbers and extend patterns. These tasks reflect how math skills evolve beyond basic enumeration.

To study this, scientists at the Technical University of Munich (TUM) and the University of Cologne developed an AI-driven eye-tracking system. The software monitors students’ gaze patterns as they solve math problems, providing real-time feedback. A webcam tracks their eye movements, and an AI algorithm translates this data into heat maps. Red areas show where students focus most, while green areas highlight places they barely glance at.

"The AI system classifies the patterns," explains Achim Lilienthal, a robotics professor at TUM. "Based on this, the software selects learning videos and exercises tailored to the student's needs."

The AI-based learning tool is designed to work with just a standard webcam and a modern computer. Previously, high-end eye trackers costing thousands of dollars were needed for this kind of research. The new system is more affordable and adaptable for classroom use.

Researchers programmed the AI to detect different problem-solving strategies. For instance, when students count dots in a ten-frame grid, some quickly recognize missing numbers and skip ahead. Others methodically count each row, indicating a need for additional support. The AI sorts students into different learning paths based on their gaze patterns, ensuring individualized instruction.

"Individually customized lessons for high-achieving children are also possible in the future," says Lilienthal. The system creates automated reports for teachers, helping them provide personalized support. Instead of relying on test scores alone, educators can now see exactly where students struggle in real time.

The Wulfen Comprehensive School in Germany is the first institution to use this AI-based learning system. A standardized math test revealed that one-third of the 180 incoming fifth graders had arithmetic difficulties. Traditionally, teachers could only give one-on-one support to a handful of students. Now, five students at a time can receive targeted instruction using the AI system.

"We are delighted that we can now support significantly more children in their basic math skills," says one school official. "This means we can help more learners improve their math performance than in the past due to a lack of teachers."

With teacher shortages and increasing class sizes, AI-driven tools offer a scalable solution. The system provides instant feedback, reducing the time needed for traditional assessments. More importantly, it ensures that struggling students receive help before falling behind.

The researchers believe this technology could be expanded beyond elementary math. Future applications may include reading comprehension, science education, and even special needs instruction. As AI continues to evolve, tools like this could redefine personalized learning worldwide.

Note: Materials provided above by The Brighter Side of News. Content may be edited for style and length.

Like these kind of feel good stories? Get The Brighter Side of News' newsletter.