Dear Students and Researchers, The Student and Outreach Subcommittee of the Information Theory Society is excited to present the IEEE ISIT 2026 Data2Information (D2I) Competition, a global student challenge at the intersection of Information Theory, Machine Learning, and Wireless Communication. This year’s competition focuses on Reflective Intelligent Surfaces (RIS), Neural Mutual Information Estimation (NMIE), and Physical Layer Security (PLS).
We know—it sounds like three different topics, but trust us: somehow, we made it one, coherent, and exciting competition.
The challenge will follow a hybrid format: participants will work on the tasks online, while final presentations will take place during a special session at ISIT 2026.
Teams may consist of up to five members. At least one team member is required to attend ISIT 2026 in person and present the results.
The challenge is organized as a sequence of increasingly complex tasks, guiding participants from data exploration to optimization and security:
Task 0 – Channel Exploration Participants work with RIS-based CSI measurements across different configurations (RIS phases, frequencies, positions, antennas).
Task 1 – Mutual Information Estimation Given CSI data, participants estimate the mutual information (MI) of RIS-assisted channels under Gaussian signaling.
Task 2 – Phase and Beamforming Optimization Participants design RIS phase configurations and transmit beamformers to maximize channel MI.
Task 3 – Secret Key Generation Given channel observations from an eavesdropper (Eve), participants attempt to reconstruct secret keys generated by legitimate users (Alice and Bob).
The overall goal is to bridge theory and practice, combining signal processing, machine learning, and information theory into a unified data-driven framework. Organized by the IEEE Information Theory Society, this competition offers undergraduate and graduate students worldwide a unique opportunity to work on a modern and challenging problem, while competing for prizes and recognition. Full details about the competition, including dataset, tasks, and evaluation, are available here:
Note that you can register as a team or individually: if you register individually, the organizers will help match you with teammates. Join the community: