Frequently Asked Questions

There are many AI training courses, and some of them are even free. Why should I consider taking courses with Beaver-Edge AI while preparing for AI Olympiads?

​

1. Some AI training courses on other platforms are at one extreme—they are too applied. They teach students how to directly use AI software libraries, such as Scikit-learn for linear regression and support vector machine models. However, these courses often treat these models as black boxes, without explaining the mechanisms behind them, such as why they work and how they are derived. AI Olympiads, however, require students to have a solid understanding of these mechanisms.

By contrast, our curriculum bridges this gap. We teach students AI models with sufficient theoretical rigor and depth.

​

2. Some AI training courses are at the opposite extreme—they are too theoretical. For instance, some universities offer AI courses that are accessible for free. However, these courses are often highly theoretical, with prerequisites in mathematics that go far beyond the level of most high school students. Additionally, lectures in these university-level AI courses tend to focus exclusively on theory, such as deriving formulas for machine learning models, and presume that students already have advanced coding skills to implement those models. This is not the case for K-12 students, as high schools typically do not teach tools like NumPy or PyTorch, which are essential for coding in AI.

​

By contrast, our curriculum bridges this gap. First, for the theoretical side of AI, we developed AI200, a course designed to teach students the mathematics needed for AI Olympiads but not covered in the K-12 curriculum or math Olympiads. Second, our curriculum emphasizes a substantial coding component. For each AI model and algorithm, such as Principal Component Analysis (PCA), Transformers, and Generative Adversarial Networks (GANs), we guide students step-by-step on how to code them from scratch. ​

​

3. If students attempt to prepare for AI Olympiads by searching for learning materials online, they will face high time and monetary costs. 

First, many students lack the ability to identify the best resources. For example, while there are numerous notes and videos online about PCA, the model itself is highly mathematical, and the concept of eigenfaces is abstract. Students may struggle to determine which resources are accurate or effective. In some cases, they may encounter materials with errors, but lack the expertise to identify them.

Even students who can identify quality resources will need to invest significant time in reading and comparing them.

​

Our curriculum eliminates these issues. When we developed our course materials, we thoroughly compared a wide range of resources for each topic, analyzed the pros and cons of different teaching methods, and presented the content in a clean and intuitive way, using our own explanations.

4. AI Olympiads are not the same as generic AI education; they have unique features, similar to how AMC/AIME/USAMO differ from general math education or how USACO differs from generic computer science education. Many AI training materials available on various platforms are designed for generic AI education, not specifically for AI Olympiads.

By contrast, our curriculum is specifically designed to help students prepare for AI Olympiads. For example, we offer lectures that teach students how to solve AI Olympiad problems. We adopt the perspective of problem writers to understand how these competitions are structured and guide students on how to answer questions effectively.

​