教AI教学：雷内·基齐尔塞克与国家辅导观测站的雄心使命

从学前教育到终身学习的创新，源于杰出的人才。

Innovation in pre k to gray learning is powered by exceptional people.

十五多年来，各类规模和阶段的教育科技公司都信赖高等教育来发掘推动变革的人才。

For over fifteen years, Edtech companies of all sizes and stages have trusted higher education to find the talent that drives impact.

当特定的技能和经验至关重要时，高等教育是能够切实交付成果的合作伙伴。

When specific skills and experiences are mission critical, higher education is a partner that delivers.

高等教育提供全职、兼职和高管招聘服务，深知您所需的市场人才。

Offering permanent, fractional, and executive recruitment, higher education knows the go to market talent you need.

了解更多，请访问 hireedu.com。

Learn more at hireedu.com.

那就是 hireedu.com。

That's hireedu.com.

仅仅打造一个我们希望有人会采用的定制化工具，其影响力永远无法与试图影响这些超大规模平台正在做的事情、并改善这些工具所体现的教学法相比，即使前者设计得更好。

Building sort of a bespoke tool that we hope some people will adopt is just never gonna have the same impact, even if it is much better designed than trying to affect some of what these hyperscalers are doing and trying to improve the sense of pedagogy that these tools are having.

如果当前的学习模式效果不佳，我认为让学生使用更符合教学实证、并能像优秀导师那样引导他们而非直接给出答案的工具，是正确的方向。

And if study mode is not working well right now, I think it's the right idea that students use something that is more aligned with pedagogical evidence and that helps them by not just giving them the answer, but by sort of talking them through like a good tutor would.

欢迎收听《教育科技内参》，本节目是覆盖教育科技行业的顶级播客。

Welcome to Edtech Insiders, the top podcast covering the education technology industry.

从融资动态到影响力，再到从幼儿教育、K12、高等教育到职场领域的AI进展，您在这里的《教育科技内参》都能找到。

From funding rounds to impact to AI developments across early childhood, k 12, higher ed, and work, you'll find it all here at Edtech Insiders.

内参。

Insiders.

要订阅本播客、查看我们的通讯以及活动日历，请访问我们的平台。

To subscribe to the pod, check out our newsletter, and also our event calendar.

如需深入了解，请订阅《教育科技内参Plus》，您将获得独家内容、加入我们的WhatsApp频道、优先参与活动，以及获取Alex和Ben的幕后见解。

And to go deeper, check out Edtech Insiders Plus, where you can get premium content, access to our WhatsApp channel, early access to events, and back channel insights from Alex and Ben.

希望您喜欢今天的节目。

Hope you enjoyed today's pod.

本周《教育科技内参》播客带来了一期极其精彩且特别的节目。

We have an incredibly exciting and really special episode this week of the Edtech Insiders podcast.

我们对话的这位人士，说真的，我认为是全国乃至全球最聪明、最致力于推动教育科技发展的学者之一。

We are talking with, I kid you not, I think one of the smartest, most dedicated academics who is trying to push Edtech forward in the country, maybe even in the world.

雷内·基齐尔塞是康奈尔大学鲍尔斯计算与信息科学学院的副教授，他领导着康奈尔大学未来学习实验室，并负责国家辅导观测站。

Rene Kizilcec is an associate professor in the Bowers College of Computing and Information Science at Cornell University where he directs the Cornell Future of Learning Lab and leads the National Tutoring Observatory.

基齐尔塞研究技术与教育中的行为、心理和计算方面，以指导促进学习、公平以及学业和职业成功的实践与政策。

Kizilcec studies behavioral, psychological, and computational aspects of technology and education to inform practices and policies that promote learning, equity, and academic and career success.

他的研究成果发表在《科学》杂志和美国国家科学院院刊上。

His work has appeared in science and the proceedings of the National Academy of Sciences.

他的工作获得了多项奖项，他拥有斯坦福大学传播学博士学位和统计学硕士学位。

It's won multiple awards, and he holds a PhD in communication and a master's in statistics from Stanford University.

雷内·基齐尔塞，欢迎来到Edtech Insiders。

Rene Kizilcec, welcome to Edtech Insiders.

谢谢你邀请我，亚历克斯。

Thank you for having me, Alex.

那我们先聊聊你的背景吧。

So let's kick off with a little bit of your background.

十多年来，你一直密切关注教育科技领域，致力于推动其建立在扎实的实证基础上，并真正实现变革性影响。

You have been paying so much attention to the Edtech world and how it can be more informed by serious evidence and how we can really make transformational impact for a decade or more.

跟我们讲讲你是如何进入教育科技领域的，以及你是如何逐步开展目前国家辅导观测站工作的。

Tell us about how you got into Edtech and some of your work leading up to your current work with the National Tutoring Observatory.

很乐意分享。

Happy to.

我觉得自己已经走了一个完整的循环。早在最初，我还是英国的一名大学生时，就曾在美國的一个夏令营里辅导孩子编程，那正是我教育生涯的起点。

I feel like I've come full circle since the very beginning when I started tutoring kids in a summer camp in The United States when I was a college student in The UK, teaching them how to program.

我当时在进行结对编程，这真正激发了我对教育和教学的兴趣。

I was pair programming, and that's really what sparked my interest in education and teaching.

我连续三个夏天都这样做，贯穿了我的整个本科阶段，之后才前往斯坦福开始我的博士研究。

I did that for three summers in a row, my entire undergraduate, before I moved to Stanford to start my PhD.

在斯坦福，我最初研究的是汽车界面，完全是另一回事，专注于人机交互问题，直到慕课突然兴起——大规模开放在线课程。

At Stanford, I first was working on car interfaces, something completely unrelated, and human computer interaction questions there, when suddenly MOOCs happened, massive open online courses.

那时正是慕课刚刚起步的时候，Coursera刚刚成立，教育科技也随后兴起，Sebastian Thrun创建了Udacity。

It was when the very beginning of them launching, Coursera just started, Edtech started shortly after, Sebastian Thrun created Udacity.

当时一群博士生聚在一起，说：这太有意思了。

And a group of students, PhD students at the time, came together and said, This is really interesting.

我们想研究这个。

We want to study this.

于是我们与斯坦福大学的教授合作，建立了当时的Linux实验室和学习分析实验室。

And we partnered with faculty at Stanford in order to create the Linux Lab, the Learning Analytics Lab at the time.

我们对这些大规模开放在线课程中的情况进行了首批分析之一。

And we did some of the first analyses of what was going on in these massive open online courses.

谁参与了这些课程？

Who was in them?

人们为什么参与这些课程？

Why are people in them?

通过率如何？

What about the success rates?

我们该如何定义这种新型非正式学习环境中的成功？

How do we define success in this kind of new space of informal learning?

借助这些早期数据，我们得以回答所有这些问题。

All of these questions we're able to answer with some of this early data.

那是一个非常令人兴奋、节奏飞快的时期，当时人们还不明白发生了什么。

It was a very exciting time, very fast moving, when people didn't understand what was going on.

在某些方面，这与我们如今对人工智能的理解和应用处境非常相似。

In some ways, it's very reminiscent of where we're with AI right now to make sense of what is going on and how it's being used.

因此，我认为这是一个类似的环境。

So I see this as the similar environment.

当然，我们就是在这里第一次见面的，亚历克斯，当时你在Coursera。

This is where we, of course, first met, Alex, when you were at Coursera.

我过去那里，介绍我们关于在线学习者动机的一些研究发现。

I was coming over there presenting some of the research findings that we had on motivation of online learners.

我们做了大量的干预研究。

And we did a lot of intervention research.

因此，当我们刚开始进行慕课研究时，我们弄清楚了谁在那里，以及他们为什么在那里。

So when we started with the MOOC research, we figured out who's there and why are they there.

我们很快意识到，很多人虽然希望完成课程，但却未能实现这个目标。

And one of the things we quickly understood is that a lot of people are there and want to finish the courses, but fall short of that goal.

因此，当时我们开展的下一组研究是围绕支持这些课程中学生的干预措施，其中一些是基于行为科学的干预，帮助学生提升规划、自我调节和元认知能力。

And so the next set of research studies we ran at the time was around interventions to support students in those courses, some of them more behavioral science interventions to help them with planning, self regulation, metacognitive skills.

另一些则更侧重于动机心理学层面，试图在那些让人感觉与他人脱节的环境中培养学生的归属感。

And then some were more on the motivational psychological side, trying to induce a sense of belonging in environments that can feel a little disconnected from other people.

有时，如果这是一门来自斯坦福或哈佛这样的大学的课程，你甚至可能感到一种身份威胁，因为你身处一个可能让你感到不自在的环境。

And sometimes if it's a course from a university like Stanford or Harvard, you might feel sort of an identity threat even being in that course, an environment that you might not feel very comfortable in.

因此，我们研究了这些问题，试图在大量不同课程中促进学生的参与度。

And so we worked on some of those questions to try and promote students engagement at scale across a lot of different courses.

之后，我短暂地前往亚利桑那州立大学，随后才来到康奈尔大学。

After that, I moved for a brief time to Arizona State University before coming to Cornell.

在那里，我们研究了亚利桑那州立大学的在线学位项目，比较这些项目中的学生与面授项目学生的学业表现，并探讨如何通过课程设计来支持学生在所谓‘真实学位项目’中的成功。

There we worked on the online degree programs at ASU and comparing how students in those programs were performing to students in the in person programs and how to make course design decisions that could support student success with quote unquote real degree programs.

在康奈尔大学，我创立了‘学习未来实验室’，研究与教育技术相关的各种议题，涵盖K-12教育、高等教育以及职业学习等领域。

And then at Cornell, I founded the Future of Learning Lab, where we study all kinds of things related to Edtech, both in K-twelve, higher ed, as well as professional learning.

其中一些早期工作聚焦于随机对照试验，尝试通过干预措施来支持学生的学习。

Some of that early work focused on RCTs, controlled trials, trying interventions to support student learning.

关于人工智能系统中的公平性和偏见及其缓解方法，已经有一整套研究工作。

There's a whole body of work on fairness and bias in AI systems and how to mitigate it.

最近，我们重点关注辅导领域，并建立研究基础设施，帮助研究人员和开发者更好地了解辅导中哪些做法是有效的。

And most recently, we've had a strong focus on tutoring and building up research infrastructure to support researchers and developers in the space to learn more about what is effective about tutoring.

这是一个令人兴奋的项目。

It's such an exciting project.

我们希望尽快参与到国家辅导观测站项目和百万辅导行动中。

We wanna get into the National Tutoring Observatory work ASAP in the Million Tutor Moves project.

但在那之前，我想强调一下您工作中的一个方面，我很想听听您谈谈：在慕课时代，您非常关注学习成果的公平性——谁完成了课程，谁没有完成，以及这些结果如何在全球范围内、有无学位者之间、或不同社会经济背景的人群中产生分化。

But before we do, there's one aspect of your work that I want to highlight, and I'd love to hear you talk about, which is that you paid a lot of attention in those MOOC days about equity of outcomes, about who was making it through those courses, who was not, how it divided around the world or with people with or without existing degrees or from different socioeconomic backgrounds.

您尝试的一些干预措施实际上开始显著缩小了这种差距。

And some of the interventions that you tried were actually starting to really narrow the gap.

但并不是总能成功。

It didn't always work.

也并非总能推广开来。

It didn't always scale.

但我认为，对学习成果公平性的关注对您的工作至关重要，对当前人工智能和辅导领域的发展也同样重要。

But I think that focus on equity of outcomes is relevant for your work, it's also very relevant for what's happening now in AI and tutoring.

我很想听听您对此谈谈。

I'd love to hear you talk a little bit about it.

当然。

Absolutely.

是的，我们在慕课中最早发现的一件事是，人们的成功程度差异非常大。

Yeah, one of the first things we figured out in MOOCs is that people are very differentially successful.

我们曾探讨过，这是不是因为动机不同？

And we looked at is that because motivations are different?

是的，动机确实有些差异，但更重要的是，这些学习环境对准备不足的人，以及来自与课程开发地文化背景差异较大的全球各地的人，效果要差得多。

And yes, they are somewhat different, but it's also the case that the environments just work much less well for people who are less prepared to do well in them, and for people who are coming from environments, from contexts around the world that were more different from the place that these courses were developed in.

因此，我们特别研究的一个差距是：来自富裕的西方工业化地区的学习者与来自全球南方、人均GDP较低、人类发展指数较低的国家的学习者之间的全球成就差距，后者的课程完成率要低得多。

And so one of the gaps in particular we studied was this global achievement gap between learners who are coming in from rich Western industrialized environments versus learners who are coming in from more of the global South, countries with lower GDP, lower human development index, where achievement rates in terms of completion of the course were just much lower.

围绕自我调节、帮助学生建立归属感的一些干预措施在某些情况下有所帮助，但并非在所有情况下都能缩小这些差距。

And some of the interventions around self regulation, supporting students with their sense of belonging, were helpful in some cases, but not all, in closing those gaps.

这体现了在多样化环境中扩大干预措施所带来的机遇与挑战。

It's a lesson in the opportunity, but also the challenge of scaling up interventions to really diverse environments.

当我们最初尝试这些干预措施以缩小部分成就差距时，我们聚焦于几门课程，这些课程是我们精心挑选的，因为我们对它们非常熟悉，清楚差距所在。

When we first tried these interventions to reduce some of these achievement gaps, we were focusing on a few courses that we so carefully selected because we knew them well, we knew where the gaps were.

当我们试图将这一做法推广开来，与我亲爱的同事贾斯汀·里奇合作，覆盖麻省理工学院、哈佛大学和斯坦福大学的所有慕课时，我们很快意识到，这些课程的设置方式差异巨大，每个环境都不同，要让干预措施在所有这些环境中稳定推广非常困难。

When we then tried to scale this up and work with my dear colleague Justin Reich, across MIT, Harvard, and Stanford, MOOCs that were all going up, we quickly realized that there's so much variation in how these courses are set up, every environment is different, and having an intervention scale reliably across all of those is very difficult.

因此，我们看到结果的差异性大大增加。

And so we saw a lot more variance in the outcomes.

但有些措施在这些课程中表现得更加一致。

But some things did work more consistently across them.

这些是从其他干预措施中获得的经验教训，未来可以尝试应用。

And those are some lessons learned from other interventions that can be tried out in the future.

我非常赞同。

I really agree.

我认为一定的结构化支持和‘最后一公里’的交付很重要，人们常说，和朋友一起上课。

And I think certain amounts of structure and of last mile delivery having they always said doing courses with a friend.

我认为这是你的一个发现。

I think this is one of your findings.

和朋友一起上课会让人更有可能完成课程。

Doing courses with a friend have made you people much more likely to complete it.

因此，在某些情况下，拥有社交网络或归属感干预措施——即将你在课程中的学习与你的生活和你期望的实际成果联系起来——产生了巨大影响。

So having that social network or the belonging interventions where you sort of connected what you're doing in a course to your life and the actual outcomes you want for yourself made a huge difference in some cases.

所以我认为你引入的这些心理因素对这个领域至关重要，当然也完全可以很好地应用到你目前所处的辅导领域。

So I think some of the psychological pieces that you were introducing to this are so key to that world, but also obviously transfer really well to the tutoring world that you're in now.

我很高兴你强调了和朋友一起上课这一点。

And I love that you highlight taking the course with a friend.

这在我们早期的动机研究中就出现了。

This was something that came out in the early motivation research that we did.

这只是众多动机中的一种：我和别人一起学习。

It was just one of the motivations that was out of many options, I'm doing this with somebody else.

那些人完成课程的比例要高得多，于是我们将这一点转化为一种社会问责干预措施：找到一个能督促你的人，或者找个人一起学习，鼓励这种行为，这在众多选项中是一项看起来更有前景的干预措施。

And those people tended to complete at much higher rates, which we then turned into an intervention around social accountability, finding a friend that can hold you accountable or somebody to take it together with, encourage that kind of behavior, which is one of the interventions that looked more promising among the set of options.

在COVID期间，当我们对学生的支持网络进行调查研究时，也发现了这一点，这是一个重要因素。

And it's something that we're seeing also during COVID when we did some survey research around students' support networks, that was an important factor.

所以即使你没有亲自在课堂上，知道班上的其他同学也很重要。

So knowing other kids in the class, even though you weren't there in person.

我仍然认为拥有这种社交网络非常关键，尤其是当学生越来越多地依赖设备来寻求帮助时。

And I still think it's a very important thing to have that social network, especially as students are turning much more to devices in order to get help.

我的办公时间比以往任何时候都更冷清。

My office hours are emptier than they ever have been.

我认为这是因为ChatGPT和其他工具的出现，学生们更倾向于去那里快速获取答案，而不是排队等候面谈，或者顶着伊萨卡的寒冷前往教学楼。

I do think that is because of ChatGPT and other tools and students going there for a quick answer rather than stand in line in front of an office or weathering the Ithaca cold to make it over to the building.

因此，建立这些社交联系非常重要，教师应积极通过团队合作、小组作业等方式鼓励学生相互交流。

So it's important to have those social connections and have instructors actively encourage that to happen through teamwork, team assignments, other kinds of ways to get students to talk to each other.

完全同意。

A 100%.

随着我深入探索人工智能领域，与更多人交流并阅读更多资料，我越来越确信，一种具有社交属性的人工智能不仅不可避免，而且可能对教育产生变革性影响。

And I'm more and more convinced as I explore the AI field and talk to more people and read more that a social version of AI is not only inevitable, but it could be transformational for education.

我认为，一个人与ChatGPT或Gemini来回互动的模式，是我希望我们能够打破的一种范式，转而让AI更多地支持人与人之间的关系和互动。

I think this idea that one on one person and chat GBT person and Gemini back and forth is a paradigm that I hope we're gonna actually start to break and have more of AI supporting human to human relationships and interactions.

我们将会看到这是否真的发生，但这是我的希望和预测。

We'll see if that truly happens, but that's my hope and prediction.

这正好引出了你正在做的国家辅导观测站项目，这项工作非常有趣。

This perfect segue to what you're doing with the National Tutoring Observatory, which is incredibly interesting work.

首先，能给我们讲讲国家辅导观测站的由来吗？你与教育科技和AI相关的具体目标是什么？

So first off, just tell us about the origins of the National Tutoring Observatory and what some of your goals are in relationship to Edtech and AI.

是的。

Yeah.

在教育研究和教育技术文献中，个性化学习显然有着悠久的历史。

We clearly have a long history of personalized learning in the education literature and the Edtech literature.

国家辅导观测站的目标是理解一对一或小组辅导在不同人群和不同情境下真正起作用的原因。

The National Tutoring Observatory has the goal of understanding what it is about tutoring one on one or small group instruction that actually makes it work for different people in different contexts.

为此，我们正在与多家辅导机构合作，收集迄今为止最大的辅导数据集——我们称之为‘百万辅导行为数据集’，并将这些数据开放给研究和开发使用，以理解辅导为何有效，并开发更好的模型，包括AI模型、大语言模型和多模态模型，来辅助完成辅导工作。

And to do that, we are working with a number of tutoring providers to collect the largest data set, which we call the million tutoring moves data set of tutoring and making that data available for research and development in order to understand what it is that makes Tutoring really effective and to have better models, AI models, LLMs, multimodal models that can help do the job of a tutor.

当然，我们的目标并不是用这个取代教师。

Now, it's important that the goal is not to replace teachers with this, of course.

我们的目标是确保当AI参与其中时，无论是以一对一的形式，还是作为团队协作的支持角色，它都能理解什么是良好的教学法。

It is to make sure that when AI is in the loop, whether it is in a one on one fashion or it is a supporting role for a teamwork, that it understands what good pedagogy looks like.

目前一个主要问题是，大型语言模型在训练过程中所收集的数据，并不包含优秀教师实际教学行为的数据，对吧？

One of the big problems right now is that the data that LMs are sort of hoovering up in order to be trained does not have any data of what good teachers are doing, right?

因为目前根本没有关于优秀教师教学行为的大规模数据集。

Because there are just no large data sets of what good teachers are doing.

因此，我们的目标是创建这样的数据，以推动学习科学的发展，同时也有助于开发更能支持学生学习和教师发展的技术。

And so our goal is to create that data to inform the science of learning, but also to inform better technology that can support student learning and teacher development.

当然，这项工作面临的一个重要挑战是隐私问题。

Now, one of the important challenges with this work is of course privacy.

我们在数据去标识化以及为这一目的公开发布哪些数据方面，极为谨慎。

And we are extremely careful about how we are de identifying data, what data will be publicly released for this purpose.

我们特别关注的是谁可以使用这些数据以及用于何种目的。

And one of the things that we pay specific attention to is who can use the data and for what purposes.

因此，国家辅导观测站正在提供这一资源，以推动科学进步和该领域的技术发展，同时保障辅导过程中学生的隐私以及担任辅导教师的 tutors 的隐私。

And so the National Tutoring Observatory is providing this resource to advance the science, to advance development in the space, while maintaining the privacy of students who are being tutored in the session and the tutors that are the teachers in these sessions as well.

我听说过这个‘百万辅导行动’以及你们在国家辅导观测站所做的工作。

I heard about this million tutoring moves initiative and what you're doing with the National Tutoring Observatory.

你们还与卡内基梅隆大学以及那里的几位顶尖教授合作，包括肯·卡廷格，他在教育研究领域绝对是传奇人物。

You're also doing it in partnership with Carnegie Mellon and some really top professors there, including Ken Cattinger, who's an absolute legend in educational research.

我一听到这个消息，就立刻觉得：这绝对是100%必要的，原因就是你们刚才说的那些。

And as soon as I heard it, I said, oh, this is a 100% needed for all the reasons you just said.

对吧？

Right?

大语言模型是基于互联网训练的。

LLMs are trained on the Internet.

它们训练所用的数据集来自书籍、新闻机构购买的海量数据，但这些数据中都没有包含成功的教育对话，也没有反映优秀教师或辅导者的行为。

They're trained on huge corpuses of data out of books, huge corpuses of data they're buying from news outlets, none of which has successful educational conversations or what good teachers are doing or what tutors are doing.

因此，许多现有的引导式学习模式和辅导工具都在试图猜测：辅导者应该怎么做？辅导者会怎么做？

And as a result, a lot of these guided learning modes, a lot of the tutoring tools that are being built are trying to sort of guess, basically, trying to figure out what should a tutor do, what would a tutor do.

但大语言模型本质上是数据库。

But LLMs are database.

这正是关键所在。

That's kind of the whole point.

它们不是基于角色的。

They're not roles based.

它们是数据库。

They're database.

因此，拥有大量成功的辅导互动数据集具有极其重要的价值。

So having that huge dataset of successful tutoring interactions is incredibly valuable.

当然，学生与教师在课堂、家庭或辅导中心进行的一对一辅导互动可能并未被记录，也无法转化为辅导数据集。

And, of course, one on one tutoring interactions that happen between a student and a teacher in a classroom or in a home setting or in a tutoring center may not be recorded, may not be able to be turned into a tutoring dataset.

但这些在线辅导机构——其中许多正是你正在合作的——实际上拥有庞大的数据集，并且具备专有数据。

But these tutoring providers that were online, many of which you are working with, actually do have huge datasets, and they have proprietary datasets.

你所做的，是将它们整合起来，提取部分数据，合并成一个庞大的辅导互动语料库，并利用这些数据来分析哪些类型的辅导和互动更有效。

What you're doing is putting them together, taking parts of their data, and combining them into a massive corpus of tutoring interactions and also using them to figure out which types of tutoring and interactions work.

因为这实际上引出了我的问题。

Because, actually, that leads to my question.

几十年来，人们一直都知道一对一辅导相比其他类型的学习具有显著效果，特别是基于掌握的学习，也就是著名的布卢姆双西格玛学习提升。

People have known for many decades that one on one tutoring has outsized effects compared to other types of learning, and mastery based tutoring, in particular, is the famous Bloom's two sigma enhancement of learning.

但辅导长期以来一直很有效。

But tutoring has been effective for a long time.

人们早就知道它有效。

It's been known to be effective.

但与此同时，他们其实并不清楚辅导中究竟是什么因素使其有效，这一点想想真是令人震惊。

At the same time, they don't actually know what about it is effective, which is kind of mind blowing when you think about it.

所以跟我们说说这个观点，我们早就知道这种学习方式有效。

So tell us about that, the idea that we've known this PNASIA for learning.

它很昂贵，但很有效。

It's expensive, but it works.

但我们实际上一直没能拆解出辅导中究竟发生了什么，从而将这些因素应用到其他项目或产品中。

But we actually haven't been able to unbundle what's actually happening in tutoring that's making that happen in a way that we can inject into other programs or products.

没错。

That's right.

令人惊讶的是，我们对导师在学习成果方面哪些行为真正有效知之甚少。

It's remarkable how little we actually know about the moves of a tutor that are effective when it comes to learning outcomes.

我们拥有大量来自新冠疫情后利用疫情救济资金实施的大型辅导项目的数据。

We have a lot of data from large tutoring programs that have been implemented in the wake of COVID with COVID relief funding.

这些数据让我们对辅导项目的有效性有了许多新的认识。

That has led to a lot of insights into just how effective tutoring programs are.

但我们并不清楚这些辅导项目中究竟是哪些具体行为——即导师实际所做的事——导致了这些效果。

But one of the things that we don't know is what exactly it is in those tutoring programs that is causing those effects, the things that the tutors are actually doing.

已有研究探讨了小组规模应该多大，也有研究分析了项目规模多大更有效。

There are studies on how large the group size should be, are studies on sort of how large of a program is more effective.

一般来说，这些项目规模越大，对每个学习者的有效性就越低。

And generally speaking, the more you scale these programs, the less effective they become for the individual learner.

我们迫切需要弄清楚这些项目中究竟隐藏着什么样的‘秘诀’，使得它们在效果良好时能够发挥作用。

And we really need to understand the secret sauce that is in these programs that makes them work well when they work well.

我们的目标正是利用这个庞大的数据集，来理解当学生在某个概念上遇到困难并说‘我对此感到困惑’或‘我不懂’时，辅导者在那一刻可以有哪些不同的回应方式。

And that is our goal, is to use this massive data set to understand if a student is struggling with a certain concept and says, I'm confused about this, or I don't know, What are the different ways that a tutor in that moment could respond to the student?

这些回应方式中，哪一种更好？

And which one of those options is better?

哪一种更有可能帮助学生更快地意识到自己的误解，并加以克服和纠正？

Which one of those is more likely to lead to the student sort of more quickly realizing what the misconception is, overcoming it, correcting it?

这正是我们希望理解的内容。

That is what we want to be able to understand.

我们将辅导对话的数据、白板数据、有时可用的视频数据、测试成绩数据、以及学生同时使用的智能辅导系统和完成的课后小测数据结合起来。

And we're connecting the data of what is happening in the tutoring sessions, the conversation with data from whiteboards, with data from the video sometimes if it is available, with data about test scores, with data about the intelligent tutoring system that they might be using at the same time and doing exercises in with the end tickets, the ticket that they sort of fill out, the exit ticket that they fill out at the end of the tutoring session.

通过整合所有这些数据，真正弄清楚辅导者所做的哪些行为能帮助学生取得进步、更好地掌握概念。

Combining all of that to really get a causal understanding of what it is that a tutor does that helps a student move forward, progress, and learn the concepts in a robust fashion.

而这一点之所以尚未实现，是因为此前缺乏相关数据。

And that is something that just hasn't been done because the data hasn't been available.

就这么简单。

It's as simple as that.

以前没有足够的数据来回答这些问题。

There wasn't enough data to be able to answer these questions before.

这阻碍了教学与教育科学的发展，因为我们无法回答这些问题。

And it has held back the science of teaching and instruction because we couldn't answer those questions.

因此，我们的目标是让所有人都能使用这些资源。

And so our hope is to be able to make this accessible to everyone.

而其中很大一部分是我们意识到，仅有数据是不够的。

And actually a big part of that is that we realized that the data alone is not good enough.

我们可以把这些数据公开出来，但数据量庞大，很少有研究人员具备处理如此大规模文本数据集的技能。

We can put this data out there and it's large and not that many researchers will have the skills of processing such large textual data sets.

有时还涉及音频数据。

Sometimes there's audio involved.

因此，我们开发了一款工具，让研究人员能够轻松标注辅导过程中的各种情况。

And so we created a tool which allows researchers to very easily annotate what is happening in the tutoring sessions.

这个工具允许教师像使用ChatGPT来提示、修改文本或其他操作一样，通过提示来查找重述、提供纠正性反馈，或任何你想要检验的理论——比如哪些做法是有效的，或者导师是否通常会这样做。

So this tool allows a teacher to say, just like you use ChatGPT to prompt, to revise your text or do other things, you can prompt this tool to find instances of revoicing or giving corrective feedback or whatever it is that you're looking for, what your theory is that you want to test about what is effective or whether tutors are doing this typically.

这个工具能帮助你在数据中找到这些实例，并简化分析，以理解其中发生了什么。

And the tool allows you to find those instances in the data and be able to simplify those analyses to understand what's going on.

因为数据集的价值取决于它被使用的程度以及使用者的能力。

Because a data set is only so good as it is being used and the abilities of the people sort of using it.

因此，让大量研究人员、社会科学家，以及希望在此基础上进行开发的开发者能够使用它，对于实现真正的变革至关重要。

And so making it accessible to a large set of researchers, social scientists, as well as developers who want to sort of build on top of it is going to be important to affect real change.

是的。

Yes.

让我们谈谈最后一点，即它对研究人员和社会科学家，以及那些试图构建这些系统的人的相关性。

Let's talk about that last piece about it being relevant for researchers and social scientists, but also for developers and people who are trying to build these systems.

因为正如你所知，我认为你比大多数人更清楚，在教育和教育技术的历史上，一个非常棘手的问题是，一些研究中非常扎实的发现往往无法传递到真正设计工具并将其规模化、融入学生日常生活中的人手中。

Because as you know, I think better than than most, one of the things that has been really tricky in the history of education and education technology is that some of the really robust findings in research just don't make their way into the hands or into the plans of people who are actually designing tools that then get scaled and become part of everyday life for many students.

研究与实践之间可能存在脱节。

The research can be divorced from the practice.

令我感到兴奋的是，你们正在进行的‘百万教学行为’计划，这个数据库显然对研究人员来说具有极高的价值。

And what is exciting to me about what you're doing with this Million Tutoring Moves initiative is, yes, this database is obviously incredibly valuable for researchers.

能够标注、发现不同的辅导行为，评估它们的影响，并识别出真正有效的方法，这是一个极为重要的问题。

The idea of being able to annotate, find different tutoring moves, find the impact of them, and identify what is actually working is a hugely important question.

同时，我们都清楚人工智能的发展有多快，人们多么迅速地将AI辅导工具产品化。

At the same time, we all know how fast AI is moving and how quickly people are productizing AI tutors, frankly.

他们不断构建并推出这些工具，努力改进它们。

They're building them and launching them constantly and trying to improve them.

因此，产品化的发展速度与研究的速度并不总能保持一致。

So the speed at which the productization is happening and the speed of research tend to not always be at the same speed.

我非常想听听你对这项国家辅导观测站研究的发现如何影响整个领域、政策领域和产品开发领域的期望，而不只是停留在学术圈内。

I'd love to hear you talk about what your sort of hopes are about how the findings that are gonna come out of this National Tutoring Observatory work might actually be able to influence the field, the policy field, the product development field, and not be captured in an academic world.

再过十年，我们可能会说：我们终于弄清楚了什么是好的辅导，但过去十年里，我们却一直做错了。

And then in ten years, we say, we finally figured out what good tutoring looks like, and for the last ten years, we've been doing it wrong.

是的。

Yeah.

我们已经想到了几种可能实现真正变革的途径。

There's a few different pathways that we have in mind for how this can affect real change.

其中之一是，理解什么是优质辅导的科学将帮助我们更聚焦于如何引导这些机器，以及它们所执行任务的设计。

And one of them is that the science of understanding what is good tutoring will help us be more focused about what it is that we prompt these machines to do, the design of what they're doing.

所有这些都可以通过更好地理解导师如何有效支持学生来获得启发。

All of those things can be informed by a better understanding of what are really effective ways that a tutor can support a student.

这是第一个途径。

That's one.

另一个途径是，人们在构建系统时利用这些数据，使其建立在真正优秀导师的行为和有效干预措施的基础上，用这些数据来改进现有的努力。

Another one is for people to use the data when they are building systems and have them grounded in what it is that actual good tutors are doing and what are effective moves, using that data in order to improve existing efforts.

我非常希望，如果我能把数据带到亚利桑那州立大学、GSV和其他地方，并说：嘿，这里有一个资源。

I would love for, you know, if I could bring the data to ASU, GSV and another place and say, Hey, here's a resource.

请你们在开发产品时，看一看、用一用这些数据，因为这会让产品变得更好。

Please, if you're building something, look at this, use this, because it's going to make the product be better.

在产品如何产生影响这个问题上，我非常务实。

And I am very realistic when it comes to how products can have impact.

事实上，全球许多学生正在使用诸如‘judging beauty’之类的工具来帮助他们学习。

It is simply true that a lot of students around the world are using tools like judging beauty and other things to help them study.

我们都看过那些报告，试图打造一个定制化的工具，希望有人会采用，但这种做法永远无法产生与影响这些超大规模平台相同的效果，即使它设计得更好；我们更应该尝试改进这些工具所体现的教学法，对吧？

We've all seen the reports building sort of a bespoke tool that we hope some people will adopt is just never going to have the same impact, even if it is much better designed than trying to affect some of what these hyperscalers are doing and trying to improve the sense of pedagogy that these tools are having, right?

如果当前的学习模式效果不佳，我认为让学生使用更符合教学实证、能帮助他们而非仅仅提供答案、而是像优秀导师那样引导他们的工具，这个方向是正确的。

And if study mode is not working well right now, I think it's the right idea that students use something that is more aligned with pedagogical evidence and that helps them by not just giving them the answer, but by sort of talking them through like a good tutor would.

因此，第三条路径是努力让这些数据对超大规模平台也具有实用性，他们正试图改进当前模型的表现，我们必须现实地认识到，数以百万计的学生正在使用这些工具，而且他们不会停止使用。

And so that third avenue really is to try and make the data useful to hyperscalers as well, that are trying to improve what models are doing right now, being realistic that millions of students are using them right now, and they're not going to stop using them.

对此我们无能为力。

There's nothing we can do about this.

没错。

Right.

我们希望这些模型在它们所做的事情上表现得更好。

We want these models to do better in what they're trying to do.

这其中一个重要部分是为整个社区建立评估标准，设定清晰的基准，让模型可以以此为目标不断改进。

And a big part of that is also setting up benchmarks for the community, having clear benchmarks that models can try and improve on.

目前教育领域可供超大规模平台和其他工具提供商衡量自身进展的基准实在太少了，我们根本不清楚这个领域的发展究竟意味着什么。

There's way too few benchmarks in education right now that hyperscalers and other tool providers can measure themselves against in order to see what does progress look like in this space.

我们需要更多这样的基准，来展示什么是优质的辅导，如何识别优质辅导，以及那些已被证实与更好结果相关并能促成更好结果的具体策略。

Having more of those benchmarks that show this is what good tutoring looks like, how well you're able to identify good tutoring, specific moves that we know are correlated and predictive of better outcomes and causing better outcomes ideally.

这些正是我们社区、NTO以及其他众多研究团队正在努力开发的内容。

Those are all things that we need more of in the community and the NTO, also other many other teams out there, research teams out there are working on developing.

完全正确。

A 100%.

我认为你提到的这些基准值得深入探讨，因为这是人工智能领域中一部分人非常重视、而另一些人却完全忽视的方面。

I think the benchmarks that you're mentioning is something worth pausing and double clicking on because this is a aspect of the AI world that is very deeply appreciated by some and sort of totally overlooked by others.

在传统大语言模型的应用和发展中，基准确实起到了关键作用。

You know, benchmarks in traditional LLM application and evolution have been really instrumental.

基本上，能够评估大语言模型在特定推理能力、行为模式或任务完成度方面的基准，已经明确推动了模型的演进，并激发了前沿实验室之间的竞争，看谁能走得更远。

Basically, you know, benchmarks that can assess a large language model against certain types of reasoning or certain types of behavior or certain types of tasks that it can complete have actually really very explicitly driven evolution and driven competition among the frontier labs for who can push it further.

每当像Gemini 3这样的大语言模型发布时，每当任何这些大模型升级时，它们都会展示自己在所有这些行业基准上的表现，而这些基准通常本质上是在评估大模型作为学生的表现如何。

And and, you you know, every time your Gemini three just came out, every time any of these large language models evolves, they show how it did across all these industry benchmarks, which are often basically evaluating the LLM on how well it almost acts as a student.

也就是它解决问题、推理、撰写复杂内容或进行创意写作的能力有多强。

It's how how well it can solve problems or reason or write really complex things or do creative writing.

我认为，缺乏针对教学法的基准，缺乏对优秀教学标准的衡量，这一点非常明确。

And I think the lack of benchmarks for pedagogy, the lack of benchmarks for what good teaching looks like has been really explicit.

我刚参加完谷歌AI教育论坛，印象深刻的是谷歌实际上已经基于一些核心学习科学原则，为自家产品建立了有意义的内部基准。

I just came back from the Google AI for Learning Forum, and I think I've been impressed by how Google has actually really has worked on creating meaningful internal benchmarks for the Google products using some core learning science principles.

他们正努力以一种能约束自身标准的方式推进，确保任何与学习相关的工作都能切实符合这些基准。

And they're trying to do that in a way that holds them to a standard and says anything we do that's learning oriented, we want to make sure it's meeting these benchmarks in a meaningful way.

但目前整个领域在这方面还非常匮乏。

But we have very little across the field right now.

不过，有一些基准正在开发中，还有一些即将发布。

And there are some in development, There are some coming out soon.

有一些由基金会资助的项目目前正在推进中。

There are some that that foundations have been funding that are in process.

但我很想听听你对这个愿景的看法：如果这些基准得以建立并落地，它们将如何推动教育领域AI的发展？

But I'd love to hear you talk about your vision of if some of these benchmarks were to be developed and were to be in place, how might it improve the development of AI for education?

是的。

Yeah.

基准在这一领域推动了大量行动，无论好坏。

Benchmarks are driving a lot of the action in this space, for better or worse.

教育领域的基准或教学基准的一个棘手之处在于，这是一个多目标问题。

And a tricky thing about benchmarks in education or benchmarks for teaching is that it's a multi objective problem.

它不仅仅是好或坏。

It's not just good or bad.

作为辅导者，你试图优化很多方面，对吧？

A lot of things you're trying to optimize for as tutor, right?

你试图保持学生的注意力。

You are trying to keep the students' attention.

你试图激励他们，让他们继续前进。

You're trying to motivate them in the area to keep going.

你希望他们如果有其他问题时能再次回来。

You want them to be able to come back if they have another question.

当然，你还希望识别出他们的错误观念并加以纠正。

You want to, of course, identify what misconception they have and overcome it.

有很多事情正在发生。

There's a lot of things that are going on.

这是一个复杂的领域，但这并不意味着我们应该轻易放弃它。

It's a complex space, which doesn't mean we should give up on it very importantly.

这意味着我们应该制定一套基准，就像其他领域在取得重大进展时所做的那样，并试图让前沿模型在这些各种指标上取得进步。

It means that we should come up with a suite of benchmarks, which is what is also done in other areas where people have made great progress, and try and hold the frontier models to making progress on these various metrics in there.

我们现在正在考虑的一件事是，由于很难为整个辅导过程设立一个基准，不如先从一个简单的基准开始，即识别辅导行为，对吧？

One of the things that we're thinking of right now, because it is hard to sort of do a benchmark for just tutoring overall, is to start off with a benchmark simply for tutoring moves and identifying tutoring moves, right?

能够正确识别出辅导者正在表扬学生、复述学生的话、给出例子、提供支架支持、建立反馈循环。

So being able to correctly identify that a tutor is praising the student, is re voicing what a student has said, is maybe giving an example, scaffolding, having a feedback loop.

所有这些行为在文献中都被证明与学生的学习效果相关，确保模型能够识别出辅导过程中这些行为的发生，这样当你指导模型更多地采用其中某一种行为时，它更有可能做到，并且在辅导学生时更有可能使用有效的辅导策略。

All of these moves that in the literature have been shown to be correlated with student learning, making sure that a model understands when that is happening in the tutoring session, so that if you are instructing a model to do more of any one of those, that it's more likely to be able to do that, and then it's more likely to be able to use effective tutoring moves as it is tutoring students.

结果发现，这些行为中，模型目前识别起来出人意料地困难，不同模型之间的表现差异也很大。

And it turns out that surprisingly, of these are much harder for a model to identify right now, and some models more than others than other moves.

举个例子，我们测试了旧版的Gemini，也许新版本已经修复了这个问题，但它在识别辅导过程中的表扬行为时，比Claude和GPT困难得多。

Just as an example, we were testing the older version of Gemini, maybe this is fixed in the new one, but it was having a much harder time identifying giving praise in these tutoring sessions than was Claude and GPT.

而这种情况在同一个提示下只是一个奇怪的异常值。

And it was just a weird outlier, given this very same prompt.

因此，能够逐一审视这些类别，确保我们在每个方面都表现良好，这就是一个基准的示例——虽然在某些方面是渐进式的，但它是我们逐步构建更多基准、以推动更好辅导与教学的起点。

And so being able to sort of look at each one of these categories, making sure that we're doing well on them is just one example of a benchmark, very incremental in some ways, but a place to start as we're building up more and more benchmarks that can help us build towards better tutoring, better pedagogy.

一个经常在这些研究型讨论中提到的例子是：不要直接给出问题的答案，而是询问学生他们的想法。

One example, I think, that has come up a lot in these sort of study mode pieces is not giving away the solution to a problem, but sort of asking a student for what do they think.

这在某种程度上是一件非常简单的事情，对吧？

And it's remarkably simple thing in some ways, right?

但这与大型语言模型通常被提示和训练的方式截然不同——后者通常被要求成为一个优秀的客服代理，快速而礼貌地提供答案。

And it's so different from what an LLM is usually prompted and trained to do, which is to be a great customer service agent and give you answers fast and politely.

但在这个情境下，你并不希望这样，对吧？

But in this case, you don't want that, right?

你希望大型语言模型能质疑学生的说法。

You want the LLM to question what the student is saying.

我的朋友贾斯汀·里奇有一句很精彩的话：好老师会质疑学生的答案，而大型语言模型却只是回答问题。

My friend Justin Reich has this great line around, a good tutor will question the answers of a student, whereas LMs, they just answer questions.

因此，学习模式某种程度上就是在努力做好这一点。

And so study mode in some ways is trying to do that well.

而衡量模型在这方面的表现如何，也能帮助我们向前推进。

And just a benchmark on how well a model does on that can also help us move forward.

我大力支持这项工作，并向斯坦福大学的团队致敬，特别是苏珊娜·洛布、瑞安·奈特以及AI领域的其他同仁，他们正在这一领域开发一系列评估标准，我认为这将对整个社区大有裨益。

So big advocate for this work and a shout out to the team at Stanford, Susanna Loeb and Ryan Knight and others at AI too are working on a set of benchmarks in this area, I think will be immensely helpful for the community.

完全同意。

Totally agree.

听你谈到提问时，我突然想到，我们最近刚和Polygents的雅诺什聊过，他的‘教LM’论文也指出，提问、获取更多背景信息，正是他平台上的导师们所做、而大语言模型通常不做的事。

It strikes me as I hear you talk about the questioning, and we just talked to Janosz from Polygents recently, and his teach LM paper also identified that asking more questions, getting more context is a core thing that that tutors in his platform do that LLMs don't tend to do.

这让我觉得特别有共鸣。

I I it just feels very resonant.

我想到一个讽刺的现象：人类历史上最早的重要AI工具之一就是Eliza，它是一个极其基础简单的AI，几乎只做一件事——不断提问，假装自己是心理治疗师。

It strikes me as just there's this irony that, you know, one of the first major AI tools in human history was Eliza, which is incredibly basic and simple AI that basically would do almost nothing but ask questions, basically pretend to be a therapist.

这让我觉得，如果Claude能稍微更像Eliza一点，它或许会成为一个更好的导师，尽管这种逻辑看起来几乎是世界上最笨的。

And it just strikes me that, like, if Claude act more a little more like Eliza, I think he would actually be a better tutor even though that is, like, almost like the dumbest logic in the world.

但如果你对一个学生说，你对这个有什么看法？

But if you're telling a student, what do you think about this?

他们回答了x y z，然后你问，你为什么这么认为x y z？

And they say x y z, and you say, why do you think that about x y z?

而不是一上来写四段关于它看法的文字，最后才问你三个问题。

Instead of four paragraphs about what it thinks about it and and then asking you three different questions at the end.

我的意思是，我觉得这样反而会更好，这想法听起来简直荒谬至极。

I mean, I think that would actually be better, which is just like the silliest thing to even think about.

但确实如此，我们在教育领域早就知道这一点了。

But it's so true, and we've known it for a long time in education.

对吧？

Right?

我们一直试图让老师少说点，让学生多说点，对吧？

We we try and get teachers to talk less, have the students talk more, right?

比如TeachFX就是一个很好的例子。

I mean, TeachFX is a wonderful example of that.

他们有一个仪表板，这是一个教师可以在课堂上使用的应用，可以录制课程，然后获得关于教学的反馈。

They have this dashboard where they show it's an app that teachers could use in their classrooms and record the session, and then they get feedback on their teaching.

他们获得反馈的一个非常有价值的内容是讲话时间。

And one of the things that they get feedback on that's incredibly valuable is talk time.

他们讲了多少， versus 学生讲了多少？他们是否让学生多说话？

How much did they talk versus did students, did they let students talk?

在他们为改善这一指标而进行的一些干预中，减少教师的讲话时间，结果显示这在课堂上产生了积极效果。

And in some interventions that they've run trying to improve on that metric, getting the teacher talk time now, they are showing that there is positive effects of doing that in classrooms.

这并不是什么高深的科学，对吧？

And it's not rocket science, right?

学生一整天都被这样教导。

Students are just taught that all day.

我们知道，这种方式并不利于学习。

We know that that is not an effective way to learn.

我们知道，主动学习更有效，互动式学习最有效。

We know that active learning is more effective, interactive learning is most effective.

因此，通过大语言模型以及在社交环境中促进这一点，将是一个需要关注的重要挑战。

And so facilitating that with LLMs, but also in social settings, is going to be an important challenge to look at.

我想回到你提到的‘超大规模平台’这个说法，因为我有一个理论：如果你取一批来自OpenAI的对话记录，比如ChatGPT的学习模式、引导式学习，或者Claude的学习模式，你会看到一些与我们这里所说的非常相似的东西。

I wanna come back to your comment about what you call hyperscalers, because I have a theory that if you were to take a bunch of transcripts from OpenAI, you know, ChatGeVita study mode or guided learning or Claude's, you know, study and learn mode, you would see something that actually looks a little bit like what we're saying here.

你会看到学生提出一个问题，或者只说一个词或简短的回答，然后大语言模型却输出一大段冗长的内容，告诉你应该怎么做、应该考虑提供练习测验或制作闪卡，接着学生再回复两个词，然后模型又开始新一轮的输出。

You'd have a a student asking a single question or saying a a one word answer or a short answer, and then this huge, you know, word vomit from the LLM about all the things that you should do differently or you should think about maybe offering a practice quiz or to make flashcards, and then the student says two more words, and then the thing goes again.

这只是我的猜测，但基于我在该领域的某些经验，这就是我的看法。

It's just a guess, but based on some experience in the field, that's my guess.

感觉我们现在正处在一个关键时刻，正如你所说，这些工具——无论是商业消费版还是学习模式——都在大规模使用。

It feels like we are at a moment when, as you said, these tools are being used, both the commercial consumer versions and the study modes, at huge scale.

像可汗学院的Conmigo这样的工具正在大规模使用。

Things like Conmigo from Khan Academy are being used at large scale.

各种AI家教已经广泛存在，包括来自传统企业的，但它们其实并没有为这种场景专门设计。

All sorts of AI tutors are out there, including from incumbents, and yet they're really not designed for that.

我想问的是，你正在为‘百万教学行为’项目构建的这个数据集，你如何看待它？

I guess my question is, this dataset that you're building for the million tutor moves project, how do you see it?

你说你们在设计时是考虑到超大规模平台的。

You say we're designing it with hyperscalers in mind.

我的意思是，未来会不会出现这样的情况：教育科技公司和大科技公司拿到这个数据集后说，好吧。

I mean, is it a possible future that Edtech companies and big tech companies can take this dataset and say, okay.

我们要重新训练所有用于辅导或教学的系统，让它从这个数据集中学习哪些方法有效、哪些无效？

We're going to retrain anything we're doing that's trying to be tutoring, that's trying to be teaching, and make it learn from this dataset about what works and what doesn't?

这是这个项目的明确目标吗？

Is that an explicit goal of the project?

跟我们说说那会是什么样子。

And tell us what that would look like.

是的。

Yes.

我们的明确目标是，我们的工作能对提升学生的学习成果和学习体验产生尽可能大的积极影响。

It is an explicit goal that the work we're doing has the largest possible positive impact in the field when it comes to improving students' learning outcomes, students' learning experiences.

如果通过与谷歌、OpenAI、Anthropic等公司合作是实现这一影响的一条可行路径，而事实上这些正是许多学生日常使用的工具，我认为这是一种非常合理的做法。

And if one path to that is to go through working with Google OpenAI and Anthroping and others, given the reality that that is what a lot of students are using day to day, I see that as being a very reasonable approach to having that impact.

话虽如此，这些公司本身并不缺乏资源来独立开展这些工作。

Now, with that said, it's not like those companies don't have enough resources to do things on their own.

但关键的区别在于，你如何策划这样一个数据集，如何仔细思考你想要优化的目标。

But the critical difference is how you curate a data set like this, how you carefully think about what outcomes you're trying to optimize.

在策划这一资源的过程中，需要投入大量思考，这将影响到学生使用这些工具时的实际效果。

There's a lot of thought that goes into curating this resource that will have impacts in how it's then being actualized, realized when students are using those tools.

而正是在这里，我们团队真正找到了自己的定位，进行这种细致的思考。

And that's where we really see our group fitting in and sort of doing that careful thought.

我们应该优化哪些目标？

What are the outcomes to optimize for?

我们如何从数据集中提取这些目标？

How do we extract them from the data set?

当然，还要极其谨慎地保护隐私。

And of course, being extremely careful about privacy.

我一开始就已经提到了隐私问题。

I mentioned privacy at the beginning.

确保学生和教师无法被识别至关重要，必须以保护隐私的方式共享这些数据，并最大限度地带来积极成果，同时尽可能减少因共享此类数据而可能产生的任何风险。

It's going to be paramount to make sure that students are not identifiable, teachers are not identifiable, to make sure that this is going to be shared in a privacy preserving manner and, you know, has maximal positive outcomes and really minimizing the risk of any harms that come from sharing data like this.

完全正确。

A 100%.

隐私保护、匿名化、去标识化和数据聚合是其中非常关键的一部分。

And that privacy, the anonymization, de identification, and aggregation is a huge part.

你的合作方有七到八个家教提供商。

I mean, you're working with seven, eight more tutoring providers.

将不同情境下、针对不同类型学习者的多种家教课程数据整合在一起，这种综合效应非常宝贵，能够形成一个强大、有意义且多样化的数据集，适用于多种不同场景。

So the combined effect of having a dataset that combines different types of tutoring sessions in different contexts with different types of learners, for different types of subjects is very, very valuable in terms of having a robust and meaningful and diversified data set that can be used in lots of different contexts.

没错。

That is right.

我想就这两点再补充一下。

Let me pick up on two things there.

目前，我们正与七家家教提供商密切合作。

One is we have currently seven tutoring providers that we're working with closely.

还有更多机构正在联系我们，希望加入这一努力并贡献力量。

We have more knocking on our door, wanting to join the effort, being interested in contributing to the effort.

而这里非常重要的一点是数据的多样性。

And a really important part here is the diversity of the data.

我们希望这个数据集能涵盖美国各地的教学场景，并最终扩展到全球，以确保我们不仅记录富裕白人社区的教学，而是覆盖美国全境乃至全世界的教学实践。

We want to have tutoring sessions that are represented around The United States and then hopefully around the world as well in this data set to make sure that we don't capture tutoring just in sort of rich white neighborhoods, but we capture tutoring across the entire United States and the rest of the world.

我们最初选择这七家，是因为它们是我们已有合作关系的合作伙伴。

We've started with these seven because they were partners that we already have relationships with.

它们对这项工作充满热情，我们未来计划逐步扩展到美国乃至全球更多的辅导机构。

They were excited about this effort, and we're looking to expand eventually to more tutoring providers in The United States, but also globally.

目前我们专注于K-12数学领域，因为该项目的资金来自盖茨基金会和查特伯格倡议。

We're also right now focused on K-twelve mathematics because of the funding for the project is coming from the Gates Foundation and the Chatterberg Initiative.

我们希望未来能扩展到其他同样迫切需要的领域，比如科学、人文和写作，这些领域同样急需辅导支持。

We want to hopefully expand to other areas of tutoring that are just as necessary in the sciences and humanities and writing, where tutoring is also much needed.

因此，我们的目标是打造一个能够支持多个领域辅导、发展与研究的资源平台，而不仅限于我们目前的范围。

So we're looking to make it a resource that is able to support tutoring, development, and research in a number of areas beyond what we have right now.

是的。

Yeah.

所以，从实际操作的角度来说，如果有人在听这个节目，并且在一家家教公司工作，或者经营一家家教公司，或者能想到印度或巴西有五家家教公司，觉得这完全契合我们的项目。

So just to get logistical here, if somebody is listening to this and works at a tutoring company or runs a tutoring company or can think of five tutoring companies in India or in Brazil that they're like, oh, this would be a perfect match.

那么，他们最好的联系方式是什么？如何帮助我们提升这项工作的知名度，同时让我们了解他们的业务？

What would be the best way for them to connect with you and help, you know, raise awareness of your work to them and their work to you?

请通过访问 nationaltutoringobservatory.org 与我们联系。

Please reach out to us at the nationaltutoringobservatory.org.

网站上提供了电子邮件地址或联系表单，您可以填写以联系我们的合作团队。

There's an email address or contact form that you can fill out to get in touch with our partnerships team.

我们可以安排一次通话，讨论您所考虑的数据类型，以及如何与我们的项目对接，并确定整合的时间表。

We'll be able to set up a call, have a conversation about the kind of data you're thinking of, how that could fit in, figure out what the timeline might be for integration.

我们非常期待与更多合作伙伴携手，让这个资源更广泛、更具代表性，真实反映全球家教的现状。

But we're very interested in working together with more partners to make this a broader, more representative resource of what tutoring looks like.

实际上，对于许多家教公司来说，他们担心的是成为第一个分享数据的那一家。

And actually, for many tutoring companies, there's a concern around being the one that shares data.

当然，我们必须遵守所有适用的法律和用户协议。

And of course, we have to respect all applicable laws and user agreements.

但事实上，这是一个由众多辅导机构组成的联盟，它们共同为这个位于学术环境中的综合性资源做出贡献，我认为这赋予了整个项目独特的风格，使其成为创新的平台，并为众多研究人员提供资源。

But the fact that it is a consortium of many tutoring providers that are all contributing to this broader resource that is housed in an academic environment is, I think, something that gives this whole project a different flavor and makes it a space for innovation and a resource to many researchers out there.

我们还拥有一个实践社区，每月都会举办一次电话会议。

We also have a community of practice, which we run a monthly call.

亚历克斯，你已经参加过几次了。

Alex, you've joined a few times already.

太棒了。

Amazing.

我们会聚在一起，讨论研究领域的最新动态。

And we come together and we discuss what's the latest in the research.

我们会分享从数据中得出的发现，分析辅导教师的行为、这些行为如何相关，以及辅导课程如何随时间演变。

We share findings from the data that we have, sort of analyzing what tutors are doing, how it correlates, how sessions evolve over time.

我们还会讨论辅导机构所面临的挑战，比如在数据共享方面的识别问题。

And we talk about some of the challenges that providers are facing, right, which can sometimes be things as, you know, around sort of the identification of data for sharing.

它可能涉及理解导师是否遵守了他们所接受的培训，以及如何根据实际辅导课程中的观察结果调整导师培训。

It can be around sort of understanding if tutors are adhering to the training that they are receiving and how to adjust tutor training based on what is observed in the actual tutoring sessions.

因此，讨论的话题非常多，这是一个充满活力的领域，

So there's a lot of topics of conversation, and it's a vibrant space of, you

聚集了志同道合的人来交流想法。

know, like minded people to to exchange ideas.

这令人难以置信地令人兴奋，而且我觉得，现在公众对人工智能在教育中的看法正处于一个犹豫不决的阶段。

It's unbelievably exciting, and it feels like, you know, at a time when I think the public perception of AI in education is really sort of on the fence right now.

我的意思是，整个领域，甚至全球范围内，都在发生着令人惊叹的进展，我甚至可以说全球范围更甚。

I mean, there's incredible stuff happening throughout the field, including around the world, especially around the world, I would even say.

但同时，人们也对屏幕、对科技巨头、对各种新举措涌入教育系统感到担忧，许多教师担心学术诚信问题。

But you also have a fear of screens, a fear of big tech, a fear of, you know, people impinging and sort of coming into the education system with new initiatives that a lot of teachers are worried about academic integrity.

目前存在不少阻力。

There's sort of a lot of headwinds right now.

而当我听到像这样一个如此谨慎、深思熟虑、专为提升辅导效果——无论是基于大语言模型还是人工辅导——而设计的项目时，

And when I hear about a project like this that is so careful, it's so thoughtful, and it's really designed to for to do nothing but improve how tutoring outcomes work, including LLM based or human.