Unprecedented technological progress in the last decades makes information theory an ever more exciting and important discipline. The modern world is swarming with information streams pervading the radio, wires, fiber optic cables, and on-chip networks. Yet we are unable to answer the most basic questions such as the impact of delay on the capacity of multiple-antenna wireless channels, or the fundamental principles of protecting computation networks from local process variation in silicon chip fabrication. As such, the main purpose of this project is to advance the state-of-the-art in the fundamental limits of delay-constrained wireless communication. Computation of the impact of delay constraint in wireless communication will allow assessments of the degree of suboptimality of currently employed systems and industry standards, and likewise shed light on novel and higher-performing wireless systems.
The progress on non-asymptotic information theory is inseparable from understanding of non-Shannon information measures and their data-processing properties on general (non-linear) graphs. The progress on this topic is expected to provide the theoretical tools required for exploration of complex information processing systems including non-communication ones, such as fault-tolerant chips and noise-resistant circuits. Advanced converse techniques and graph-based data-processing will open information theory to new fields and is expected to reinvigorate the progress in the converse bounds for multi-terminal (network) problems.
The curriculum will be broadly disseminated through online resources, OpenCourseWare and MITx/edX. The analysis of real-world communication systems also presents a rich field for undergraduate research opportunities (UROPs). Popularizing finite blocklength results is likewise expected to have industrial impact, especially in areas related to wireless and time-critical communication. The compiled performance charts and delay-constrained analysis will guide the design of next-generation mobile standards and help in fair assessment of intellectual property.
