脱口秀大会
教育AI探索集体转向:从“答案生成”到“可视化+互动化”_我的网站

一 |

Airborne troops from China and Belarus take part in the opening ceremony for the Swift Eagle 2026 joint training exercise at a Chinese People's Liberation Army (PLA) Air Force training center on August 12, 2026. Photo: Screenshot from PLA Air Force Sina Weibo accountThe Swift Eagle 2026 joint training for airborne troops from China and Belarus commenced on Wednesday with an opening ceremony at a Chinese People's Liberation Army (PLA) Air Force training center, according to the Xinhua News Agency.
Focusing on "joint counter-terrorism operations in urban terrain," the training will focus on subjects such as surveillance and counter-surveillance, seizure-and-control and protection, and elimination and defense, Xinhua reported.
In addition, the two sides will organize seminars, friendly football matches and cultural exchange activities during the joint training, said the report.
As the fourth iteration, the joint training will help enhance the combat capabilities of the participating troops and deepen substantive cooperation between the two militaries, said Xinhua.
While details on the equipment deployed by both sides have not yet been disclosed, Wang Yanan, editor-in-chief of Beijing-based Aerospace Knowledge magazine, told the Global Times on Thursday that Belarus is likely to follow its traditional practices, focusing on low-altitude small-scale airborne operations, possibly using Il-76 large transport aircraft, while China may deploy Y-9 or Y-20 transport aircraft for low-altitude delivery.
The joint training is not a one-way learning exercise, but rather a process of building a cooperative foundation and familiarizing both sides with each other's operational habits in the counter-terrorism domain, Wang said.
While China has made significant progress in counter-terrorism and maintaining stability in recent years, the expert said, airborne troops have been rarely used in counter-terrorism operations, as local police and armed police forces are generally sufficient to handle such situations.
Airborne troops are better suited for special counter-terrorism scenarios, such as large-scale counter-terrorism operations or cases where terrorists have occupied urban building complexes and need to be forcibly removed. This joint training likely simulates such scenarios, Wang said.
The previous iteration of the Swift Eagle series airborne joint counter-terrorism training between China and Belarus took place in 2015.
In June 2015, a PLA Air Force spokesperson announced that China had dispatched an airborne special operations detachment to Belarus to participate in the Swift Eagle 2015 joint counter-terrorism training.
According to China Military Online, the training followed the 2011 and 2012 joint exercises between the two countries' airborne forces, again focusing on joint counter-terrorism operations.
The spokesperson noted that Belarus is a traditional friendly country to China.
Conducting joint airborne training with Belarus helps consolidate traditional bilateral friendship, deepen practical exchanges and cooperation between the two militaries, and enrich the content of the comprehensive strategic partnership between the two countries.
Wang noted that Belarus, located to China's west, and China share a common long-term interest in discussing counter-terrorism.
After a hiatus of several years, the resumption of the Swift Eagle joint training provides an opportunity to promote in-depth cooperation in defense and mutual trust, serving as a point of common interest for both sides.
Wang noted that the PLA's domestic training exercises have undergone profound changes in recent years.
For example, the documentary series
Zhìshèng (
Victory) broadcast in early August, showed training conducted with a no-script model, with both sides engaging in free-form confrontations and evaluating results based on actual combat performance, promoting the integration of personnel and equipment systems to generate new operational capabilities. This model has already produced notable results.
More broadly, Wang said, recent joint exercises between the PLA Air Force and foreign militaries have yielded multiple benefits, primarily on two levels.
First, they lay the groundwork for potential future joint operations, whether in counter-terrorism, maintaining stability or major international security issues, as mutual familiarity is essential for effective cooperation.
Second, joint exercises themselves contribute to regional stability. When friendly countries in the region reach consensus on security issues and conduct regular military cooperation, it has a stabilizing effect on regional security.
Overall, the PLA is building a more open and flexible military cooperation framework through both enhanced domestic combat-oriented training and expanded international joint exercises, Wang said.
。 让抽象知识变得可见,让学习过程变得可互动。
来源|多知
作者|王上
过去几年,AI进入教育领域后,最先落地的场景是“解题”。

二 |
学生拍照上传一道题,AI快速生成答案、步骤和解析。这类产品解决了学习中的即时需求,也推动了教育AI的普及。
但随着AI应用深入,一个问题逐渐显现:知道答案,并不等于真正理解。
尤其在数学学习中,很多困难并不是不会计算,而是不理解公式背后的逻辑。

三 | 函数为什么变化?
几何图形为什么这样转换?概率模型为什么成立?这些问题,仅靠文字解析很难讲清楚。
随着多模态AI能力不断提升,越来越多教育产品开始探索另一种路径:让抽象知识变得可见,让学习过程变得可互动。

四 | 可视化并不是新概念,但在过去,它更多依赖人工制作和固定内容,而多模态AI技术的发展,让动态教学内容可以被快速生成,并根据学生反馈实时调整。
从目前市场来看,无论是面向学生和家庭的消费端产品,还是服务学校和教师的教育产品,不论是大模型公司还是创业公司,都在探索如何利用AI生成动态内容、交互场景和个性化讲解。多知发现,诸多获得融资的教育AI项目,均是类似方向。

五 | 其中,数学成为最典型的应用方向。
01 当数学不再只是公式:AI尝试让知识“动起来”
数学天然适合可视化。
传统课堂中,教师可以通过画图、举例帮助学生理解,但受限于时间和工具,很多复杂过程难以动态展示。
AI的出现,让这种表达方式有了新的可能。
在消费端,AI数学产品正在从过去的“解题工具”,转向“数学探索工具”。
美国教育科技公司Brilliant旗下的Koji,就是这一方向的代表产品之一。
它通过互动动画、图形化展示和探索式练习,将函数、几何、概率等抽象知识转化为可操作的学习过程。

六 |
例如,在学习函数变化时,学生可以通过调整参数,直接观察曲线如何变化,再结合AI的引导式提问理解背后的数学规律。
相比传统刷题产品,Koji更关注的是“为什么”,而不是单纯告诉学生“答案是什么”。
类似探索也出现在其他AI数学产品中。
Mathleaks旗下MathVizy主要聚焦高中数学中的几何、函数和三角函数等内容,通过动态图像、几何模型和数学概念演示,帮助学生理解空间关系和抽象概念。
国内产品析析AI也在尝试利用AI生成动画讲解和动态过程展示,提高数学学习的直观性。用户可以在网页端、手机端或Pad端进行学习,一边在右侧的画板区域进行书写、演算与思考,一边在左侧获得AI对题目与解题过程的结构化解析。其核心交互形态之一是“墨流画板”,支持手写输入与过程推导,使得学习过程可视化,也更易于理解。
此外,Videotutor则探索了另一种方向:利用AI快速生成短视频课程。用户输入需求后,系统可以生成课程脚本、动画效果、旁白,甚至AI Tutor可以挑选自己喜欢的角色,让知识讲解更接近真人教学体验。
这些产品虽然路径不同,但背后逻辑相似:
第一,从关注答案转向关注理解过程。

七 |
过去教育AI产品主要解决“这道题怎么做”,而新一代产品希望解释“为什么这样做”。
动态图形、步骤拆解和空间变化,让学生能够看到数学推理过程。

八 |
第二,将抽象知识转化为可感知体验。

九 |
对于函数、几何、概率等内容,学生往往需要建立直觉理解,而不是单纯记忆公式。
可视化降低了理解门槛,也改变了数学学习方式。
第三,让学习从单向接受变成主动探索。
AI不只是回答问题,而是在互动过程中引导学生发现规律。
不过,这类产品也面临挑战:如何避免“视觉效果大于教学效果”,如何真正提升学习效率,而不仅仅是让内容更有趣,是商业化过程中需要解决的问题。
02 AI正在改变教师使用教学资源的方式
如果说To C产品关注的是学生如何学习,那么学校场景更关注教师如何更高效地教学。
在学校端,教育AI产品更多承担辅助教师备课、制作教学内容和开展课堂互动的角色,且不限于数理科目。
国内已有不少产品开始探索这一方向。

十 |
例如,学而思推出的九章爱学老师版,围绕教师教学全流程提供智能辅助,包括备课、批改、组卷和课堂互动等功能。
其价值并不只是替代部分重复工作,而是帮助教师减少机械性任务,将更多时间投入到教学设计和学生反馈中。
猿辅导飞象星球旗下飞象老师,则探索利用AI降低互动课件制作门槛。
教师只需要输入教学想法,AI即可生成交互课件和教学动画。过去需要专业工具制作的动态几何模型、函数变化过程,如今可以通过AI快速完成。

十一 |
海外市场也出现类似探索。
英国Virtual Class尝试将AI与互动白板结合,让学生直接在电子白板上进行数学操作。系统记录学生作图和修改过程,并形成数据反馈,帮助教师了解学生理解情况。

十二 |
从这些探索来看,AI在学校场景中的价值主要体现在三个方面:
一是降低优质教学资源制作成本。
过去制作一个动态数学课件,需要教师投入大量时间学习工具,而AI可以帮助快速生成相关素材。
二是增强课堂表达能力。
数学课堂长期依赖黑板、PPT和静态图片,而动态演示让复杂变化过程更加直观。
三是支持更加精准的教学反馈。
AI可以通过记录学生学习过程,帮助教师发现学生在哪些环节存在困难。
03 AI正在逐渐进入学习更深层的环节
从目前市场探索来看,教育AI正在发生一个明显变化:行业关注的重点,正在从“如何帮助学生快速解决一道题”,转向“如何帮助学生真正理解数学”,甚至怎样能够引导学生自主探索。
过去,AI更多解决的是“不会做怎么办”。但像数学、物理、化学这类科目的核心难点,并不只是答案和步骤,而是公式背后的逻辑、概念之间的联系,以及学生建立数理思维的过程。
可视化和互动化,正成为AI探索这一问题的重要路径。
数学、物理、化学这类科目本身具有高度抽象性,而多模态AI的发展,让文字、公式、图形、动画和交互能够结合起来,使知识不再只是停留在纸面符号上,而是以更加直观、动态的方式呈现。
这也意味着,未来数理化类AI产品的价值,可能不只是生成一道题的解析,而是围绕学生的理解过程,提供更加完整的学习体验。

十三 |
从讲清楚一道题,到帮助学生发现规律;从生成教学内容,到辅助教师设计课堂;从单向输出答案,到陪伴学生探索知识,AI正在逐渐进入学习更深层的环节。
不止是创业公司,大模型公司也在集体押注可视化+互动化,例如ChatGPT、Gemini、豆包、百度等,各个选手殊途同归。

十四 | 可以看到,基于探索、深度互动与个性化反馈的学习模式,正成为主流趋势。

十五 |
(Gemini 的交互式模拟生成功能)
这不意味着简单地增加动画和互动效果,而是需要将技术能力与教学逻辑结合起来,让视觉呈现或互动服务于认知理解,而不是成为额外的信息负担。
某种意义上来看,教育AI的探索并不是让机器替代学习,而是让学习重新回到探索和思考本身。
作者:王上。

十六 |
Current article:http://www.shangfochengcongshuaizhuang.buzz/c1vm/aps.doc
Published on:17:18:44