Past Competition / 2026
IAI²O 2026
Global Final · September 24–27, 2026
Competition design
2026 IAI²O AI Innovators Challenge Final — Problem Design Philosophy
The 2026 IAI²O AI Innovators Challenge Final was designed to assess both fundamental understanding of artificial intelligence and the ability to apply that understanding to real-world machine learning problems.
Part 1
AI Theoretical Foundations
Evaluates contestants’ understanding of core ideas across classical machine learning, deep learning, optimization, reinforcement learning, recommender systems, and modern attention-based architectures. Rather than emphasizing memorization, the problems test mathematical reasoning, conceptual understanding, and the ability to connect model equations with their practical meaning.
Part 2
Applied Machine Learning
Asks contestants to work directly with realistic datasets and develop complete machine learning solutions. The tasks require participants to make decisions about data processing, model selection, validation, optimization, and evaluation, reflecting the workflow of real AI research and development.
Together, the two parts are intended to measure a broad range of AI ability: understanding how models work, reasoning about why they work, and using them effectively in practice. The overall design rewards not only technical knowledge, but also analytical thinking, experimentation, and sound scientific judgment.
Prepared by the 2026 IAI²O Scientific Committee
Design document
Problem Design and Rationale
Prepared by the 2026 IAI²O Scientific Committee
Part 1 · AI Theoretical Foundations
Part 1 Questions
4 theoretical problems · 120 minutes · 40 marks
Part 2 · Applied Machine Learning
Part 2 Problems
Two applied problems make up Part 2 of the AIIC Final.