About
What this is
Hexa Scholar is a set of interactive lessons in university mathematics, statistics and quantitative finance, written for US college students. Each lesson takes one result that is hard to picture. You make a guess or a move first, watch the figure respond, build the proof one line at a time, and then solve practice problems whose answers are checked.
The site has two halves, like the hexagon on the home page. Math covers foundations (calculus, linear algebra, and proof, logic and sets), analysis and pure mathematics, and applied mathematics and statistics. Quant covers finance foundations, financial mathematics, and financial engineering and risk management. The map shows how the subjects connect across the two halves, for example where countable sets reappear in probability and martingales reappear in option pricing.
Every lesson follows the same five steps: Start (a big question and a short warm-up), See, Try, Prove and Apply. The Math and Quant pages list the lessons that are open now and the ones still in preparation. New lessons are added regularly.
Accounts, member features and lecture videos are planned. For now the pages store your answers in your browser only. See the Privacy Policy.
How the lessons are made
The text, the figures and the code are written with the help of AI tools. Claude (Anthropic) drafts each lesson and GPT (OpenAI) reviews it; the draft is revised until the review finds no remaining errors. No AI runs while you use the site, and nothing on a page is generated while you read it.
We do not copy textbook text, figures or exercises, and we do not reword them. Figures are drawn by code. Each lesson lists a few books for further reading.
How the lessons are checked
- Scripts. Before a lesson is published, two sets of scripts check every number it shows and the answers to its practice problems. One set runs the lesson's own code in Node.js. The other recomputes the same values independently in Python. The "Numerically checked" badge refers to these scripts.
- Cross-review. A second AI model reviews each lesson for mathematical errors, unstated assumptions, wrong answer options, and explanations that claim more than they should. The "Cross-reviewed" badge records the date this review passed.
- Browser check. Each lesson is also tried in a real browser at phone width, with a mouse and with the keyboard alone.
- Human review. No outside mathematician or practitioner has reviewed these lessons yet. We plan to ask subject-matter reviewers to check the core lessons, and we will name the reviewer and the date on the page when that happens.
The badges on each lesson say which of these checks have been done; a pending badge means that check has not happened. Agreement between AI models is no substitute for review by a person, and we do not present it as one.
What we have not checked
Passing the scripts does not make a proof correct, and two AI models can miss the same mistake. The lessons are teaching material and have not been through peer review. If something looks wrong, please tell us. See Corrections.
Who runs it
The site is run from the Republic of Korea.
Contact: [email protected]