aifollow.news
返回 量子位 AI 报道
量子位 AI 报道· · 原发布时间 AI 评分47

网易有道上线互动式数学讲题 Agent“欧拉讲题”

AI 辅助摘要

网易有道推出的“欧拉讲题”已在应用市场上线。学生可查看分步骤讲解,并针对具体步骤继续追问;产品还提供动态板书和图形演示,辅助理解几何、函数等题目的推导过程。

正文 · AI 翻译

NetEase Youdao launches interactive math problem-solving Agent “Euler Explains Problems,” now available in app stores

A child reads the answer to a geometry problem but still does not know why the auxiliary line is drawn that way; after reading the explanation of a function problem several times, the child still has no idea where to start when a condition changes. For many students and parents, the most frustrating part of learning math is not finding an answer, but having the answer right in front of them and still not understanding the reasoning behind it.

NetEase Youdao recently launched “Euler Explains Problems,” an interactive math problem-solving Agent, and it is now available in app stores. Unlike traditional problem-search products that provide answers and written explanations, Euler Explains Problems combines step-by-step explanations, interactive follow-up questions and dynamic boardwork. Students can ask questions as they follow an explanation and use visual demonstrations to understand abstract mathematical relationships.

From “telling students the answer” to “helping students understand the answer,” NetEase Youdao is taking an early lead in interactive problem explanation.

Explain clearly, answer questions thoroughly, and make it visual: How Euler Explains Problems breaks down a math problem

More specifically, Euler Explains Problems centers on three stages.

The first is step-by-step explanation: it provides the answer while also explaining the reasoning.

Euler Explains Problems does not simply list a standard answer. It first suggests a direction for thinking, then explains the problem step by step, showing the key deductions and calculations. After the explanation, the AI also categorizes the method used and identifies the core concept the problem tests.

This means students see more than how to solve that one problem. They can also understand why to begin with a particular approach, how the key steps relate to one another, and how to think about similar problems.

The second is personalized, interactive follow-up questions, allowing the explanation to follow students’ questions.

Understanding a math problem is often not a linear process. A student may suddenly get stuck at one step or develop a new question after grasping a concept.

Traditional fixed explanations generally follow a set sequence, while Euler Explains Problems lets students ask further questions about specific steps.

The change in this form of interaction is that explanations are no longer determined entirely by a preset process; they can adjust dynamically to what students actually understand.

Finally, boardwork and diagrams make abstract mathematical relationships easier to see.

Many concepts in math are highly abstract. For geometry, functions and formula derivations in particular, words alone can make it difficult for students to build a complete understanding.

Through dynamic boardwork, visual demonstrations and formula derivations, Euler Explains Problems turns abstract mathematical relationships into more intuitive visual representations.

For example, in a geometry problem, students may know they need to add an auxiliary line but not understand why it must be drawn in a particular place. Euler Explains Problems can use visual demonstrations to show, step by step, how the auxiliary line relates to the given conditions, helping students understand how the key to solving the problem emerges. For function problems, the AI can show how formulas, parameters and function graphs relate as the graph changes. For derivation problems, it uses boardwork that unfolds step by step to show changes in formulas and key calculation steps.

This approach attempts to address an easily overlooked issue with traditional explanations: students need to be told what to do at each step and also shown why.

By combining step-by-step explanations, personalized interactive follow-up questions, and boardwork and diagram-based demonstrations, Euler Explains Problems aims to create a relatively complete learning cycle.

AI does more than provide answers: NetEase Youdao takes an early lead in the next generation of learning Agents

Euler Explains Problems is intended to be more than another simple problem-search tool; it aims to be a learning Agent focused on the process of teaching math.

Math learning varies considerably from person to person: some do not understand formulas, some get stuck on a step in solving a problem, and others struggle to see how concepts connect. Compared with standardized answers, interactive explanations can adjust their approach to students’ specific questions and, through follow-up questions, guidance and checks, help students move from “understanding the answer” to “learning how to solve problems.”

Euler Explains Problems did not emerge from nowhere. It builds on nearly 20 years of NetEase Youdao’s experience in learning scenarios and its AI research and development.

In learning applications, Youdao has continued to develop intelligent learning products since its founding in 2006. After the first-generation AI Q&A Pen went on sale in February 2025, it sold out for a time, with middle-school science problems accounting for 60% of requests for help. Since then, Youdao has continued to explore a more natural AI problem-explanation experience through the Q&A Pen Space X’s “boardwork-style video answers” and the latest Q&A Pen 3’s natural-language interaction, multi-turn dialogue and contextual memory capabilities.

On the AI technology side, Youdao released its “Ziyue” model in 2023, and it became one of the first models to pass the national registration process. In 2025, Youdao open-sourced its “Ziyue-o1” reasoning model, continuing its development and open-sourcing of math models. The “Ziyue 4” multimodal model, upgraded this year, further improved its ability to handle complex visual mathematical and scientific problems.

From mathematical reasoning in its underlying models to dynamic boardwork, interactive follow-up questions and interactive learning in its products, Youdao is trying to turn its AI capabilities into a product experience that genuinely serves the learning process.

For students, the sign that they have truly learned from a problem has never been understanding the answer; it is being able to solve a similar problem on their own the next time they encounter one.

If there’s a math problem you can’t figure out, turn to Euler Explains Problems.

This article was provided by NetEase Youdao and republished by QbitAI with permission. The views belong to the original author.

发现内容有误?提交纠错