Duolingo is a leading educational technology company specializing in app-based learning. He is also a pioneer in the field of artificial intelligence (AI) for education, recently partnering with OpenAI to become one of the first companies to bring his latest language model, GPT-4, to customers. rice field.
Amazing ways Duolingo uses AI and GPT-4
adobe stock
OpenAI’s GPT technology has recently become familiar to millions of people thanks to the ChatGPT app. The ChatGPT app is said to have built the fastest growing user base ever. GPT-4 has only recently been used by a handful of applications, including the newly released paid version of ChatGPT, the latest version of Microsoft’s Bing search engine, and, as of this month, the latest version of Duolingo’s own language platform. It is now possible. – Duolingo Max.
However, this is far from the beginning of the company’s AI journey. Since its founding in 2011, its mission has been to create a “virtual language teacher” that makes language learning accessible to anyone, anywhere, anytime.
So let’s take a look at some of the ways AI has been used over the years and some of the latest AI-powered services enabled by GPT-4.
Personalized Learning, Automated Feedback, English Tests
Duolingo’s mission is to create the ultimate virtual language tutor. In order to achieve this, we must be able to serve all of our 500+ million students individually. This is one of the first applications of AI within the organization. Its students complete about 1.5 billion exercises on the platform every day, and all that data is captured and analyzed using machine learning (ML) algorithms to understand how people learn and the app. used to understand how you can use Improve how you absorb new information.
Klinton Bicknell, Head of AI at Duolingo, said:
“We can then use that data through machine learning models to figure out how to better teach and keep people engaged.”
Another important use case for Duolingo revolves around creating interactive learning experiences. An example of this is automated feedback for speech and grammar exercises that students are asked to complete. This allows app platform users to get real-time feedback on how they use the language and tips on how to improve it.
AI is also widely used in the company’s English-language testing systems, which are used by 2,000 institutions, including universities, colleges and government agencies, to assess the proficiency of non-native speakers. Here, AI is used during a remotely administered test run via webcam to detect various methods of cheating a candidate may attempt. It also automatically creates test items, assesses test difficulty, automatically scores test-takers, and verifies overall grades before test scores are published.
How is GPT-4 used?
Our partnership with OpenAI makes Duolingo one of the first third-party organizations to officially license and release products powered by GPT-4.
GPT-4 is said to be much more powerful and sophisticated than GPT-3, thanks to ChatGPT. It is reported to have been trained with 1 trillion parameters (information), compared to 175 billion for GPT-3. This means that we can provide more detailed, accurate and fluent answers to the questions asked.
Bicknell says: In the past…was the technology simply not there?”
Bicknell’s team settled on two initial use cases that were made available to users of the app.
The first is a role-playing feature that allows users to interact with AI-powered personas and perform a variety of learning-focused tasks. An example is talking to an AI barista in a cafe in Paris. Users can order drinks, specify their preferred preparation, and chat with the barista on a variety of topics. To make the experience more engaging, the characters are designed as fully fleshed out personas with their own personalities and backstories that learners can get to know through conversation and interaction.
A second GPT-4 capable feature already available is called Explain My Answer. This allows users to receive explanations on why they decide whether an answer to a question is right or wrong, receiving detailed and personalized explanations as if they were sitting and talking with a human tutor. increase.
Bicknell said: people’s language learning. ”
Working with OpenAI
Developing these features included working closely with OpenAI’s AI engineers, data engineers, and software developers. In fact, something like a symbiotic relationship quickly emerged from working practices. Both companies provided feedback on use cases and operations used to mutually improve product operations and user experience.
Bicknell says:[Explain My Answer] It was put together in one day.
“The difficult part, and a much bigger part of the work, is that the basic prototype often goes wrong in many ways. A lot of work has been done to understand how it can be reduced.”
This involved generating and labeling a large amount of data, which was used to refine the prompts (questions) sent to LLM. This data was then used to refine the pre-release version of his GPT-4 model used by the Duolino team.
We also found uses for Duolingo’s existing first-party AI algorithms. This is mainly to ensure that conversations with his GPT-4 powered bots, such as AI ballistas, go smoothly.
As we saw with Microsoft’s Bing when it was first unveiled, the first glimpse into the power (and sometimes unpredictability) of GPT-4 engulfed users in an unexpected and sometimes desirable way. I quickly found a way to produce unsatisfactory results.
To prevent your character from being destroyed in this way, you use AI routines to, broadly speaking, keep the dialogue focused on relevant topics. If not, you can gently navigate back towards the relevant task.
“There are a lot of AI models that are working, but are they working?” says Bicknell.
The aim is to provide an immersive conversational experience while keeping users safe and engaged in learning activities.
The Future of AI in Duolingo
The success of Duolingo’s AI projects to date, including its latest achievement at GPT-4, continues to make AI a priority when it comes to growing the business, creating new user experiences, and achieving its mission to be the world’s personal language teacher. It means that
New implementations on the horizon include leveraging AI and generating course content. “People are always asking for more courses and wanting them to cover more advanced topics,” he says.
“This has been a constraint in the past…We want to make sure they are achieving the highest standards of learning.It is a very time consuming and expensive process.”
Generating new course content has traditionally been a bottleneck. This is one task in particular where current generation language models have proven to be very efficient.
“They are creative and help generate a lot of exercises, so we are looking at incorporating that into our workflow,” says Bicknell.
Work is also underway to integrate GPT-4 into the company’s English language testing system. Specifically, research is underway on how to determine test takers’ language proficiency from just a few sample sentences.
And of course, delving deeper into personalization possibilities remains a top priority. The ultimate goal is what can be learned from the user’s response and how it can be used to keep that particular user engaged and satisfied that they are receiving a valuable learning experience. It’s about becoming more and more aware of what you can get.
Overall, we believe this strategy will provide invaluable insight into how AI can become a critical aspect of education systems around the world, both in online and classroom-based learning. Understanding how personalization, engagement, and experience affect our ability to teach and learn in any setting is truly transformative, making learning more effective, rewarding, and It may make it more accessible for future generations.
you can click here Watch a webinar with Klinton Bicknell, Head of AI at Duolingo, for a deep dive into the potential impact of AI on the future of education.
Follow me please twitter or LinkedIn. check out My website or some of my other works can be found here.
