Full details can be found here .
Several classical information theoretical results follow from considering sequences of infinite duration. While this assumption is adequate for systems using long error-correcting codes or long compression sequences, it hinders the relevance of information theory in other problems which require sequences of moderate length, such as low-latency communications, signal processing under delay constraints, or control of dynamical systems. It is then not surprising that the field of finite-length information theory, which considers the impact of using sequences of finite duration, has recently gained a great deal of attention. The final goal of this line of research is to establish fundamental limits and to develop strategies to attain them in the regime of finite coding delay.
Possible topics include, but are not limited to:
- One-shot information theory and information spectrum methods.
- Nonasymptotic performance bounds for point-to-point and multiterminal communication systems.
- Refined asymptotics: error exponents, dispersion, and moderate deviations analysis.
- Error-correcting codes: design guidelines and performance analysis in the finite-length regime.
- Lossless and lossy data compression at finite blocklengths.
- Delay-constrained joint source-channel coding.
- Exploiting channel feedback in code design to improve complexity–delay–reliability tradeoffs.
- Receiver design: constellation, quantization, and iterative decoding.
- Information theory for the control of dynamical systems.
Dr. Gonzalo Vazquez
Dr. Victoria Kostina