Skip to main content

Context- and Task-Oriented Communication

Periodic Reporting for period 3 - CTO Com (Context- and Task-Oriented Communication)

Reporting period: 2020-05-01 to 2021-10-31

New infrastructure in 5G networks, such as cloud-RANs, caches, relay-drones, and massive MIMO arrays, bring enormous improvements in data rates and delays even when applied naively using standard coding techniques. However, to satisfy future demands, new communication techniques perfectly tailored to exploit such infrastructure will be essential. In fact, the communication scenarios induced by these new elements differ from traditional scenarios significantly with respect to the terminals' contexts. For example, in networks with caches (i.e. additional storage spaces), terminals can download parts of their desired data directly from closeby caches. On the other hand, relay-drones can reinforce a transmit signal so as to provide a terminal with a stronger second observation of a desired signal. Finally, cloud-RANs allow terminals to obtain information about other transmit and receive signals. There is thus a basic need for highly improved context-oriented communication techniques. This need is even
more immediate in view of the rapidly increasing number of distributed decision and control systems, where the various nodes measure highly correlated signals, and thus have a priori side-information about other nodes' signals. A second major difference between communication in distributed decision and control systems and traditional communication scenarios is their task. In these systems the final task is no more to convey and reconstruct sequences of data bits or observed signals, but to make distributed decisions or take distributed actions that attain a common goal. The traditional approach uses standard LDPC, Polar, or Turbo codes to exchange scalar- or vector-quantizations of the signals measured at the various nodes, and then runs decision, control, or prediction algorithms locally based on all the accumulated information. This approach can be highly suboptimal. In particular for situations where the decisions take value in a small range, the approach can lead to huge amounts of unnecessarily transmitted data. With new, task-oriented communication techniques we aim to drastically reduce this overhead.

The goal of this project is to devise improved context- and task-oriented coding techniques for timely and important applications such as:
- Cache-aided networks.
- Optical wireless systems.
- Networks with mixed-delay traffics such as 5G.
- Coordination of autonomous agents and joint communication and sensing tasks as required for autonomous driving.
- Distributed hypothesis testing systems as encountered in the Internet of Things (IoT).
- Distributed computing systems as used for running heavy machine learning applications.
- Age of Information systems as encountered in time-critical communication systems.
Below we describe some of our advances on context- and task-oriented communcation systems that we achieved in our ERC project CTO Com.


Cache-Aided Systems
----------------------------
A large part of our internet consumption concerns only few popular videos and newspapers that can well be predicted in function of the geographical area and the time of the day. Putting in place a system of area-wide distributed caches where popular contents can be pre-stored, seems necessary to satisfy future demands in rates, delays, and energy consumption. Previous results showed that in the presence of cache memories, smart placement and coding techiques can yield significant performance gains. We extended these works to the more practical scenario with noisy communication channels and we were the first group to show that further gains are possible by employing techniques from joint source-channel coding. The gains even apply under additional secrecy requirements.



Distributed Hypothesis Testing/Distributed Decision Systems
-------------------------------------------------------------------------
Distributed decision systems are one of the main building blocks of the Internet of Things (IoT) and find applications in a wide range of domains such as health monitoring, infrastructure maintanance, natural catastrophy alarm systems, etc.
We were the first group to provide an information-theoretic analysis of distributed hypothesis testing systems with arbitrary memoryless communication channels. We proved in particular that the hypothesis testing performance of such systems not only depends on the capacity of the communication channels, as preliminary results by other authors suggested, but on the precise transition law of the channel. Moreover, employing unequal-error protection (UEP) mechanisms allows to achieve a significant gain in the testing-performance. Other significant gains can be attained through an adaptive usage of the resources, in case the constraint is on the expected usage of the resources (bandwidth) and not on the maximum usage.


Distributed Computing Systems
--------------------------------------
Different computing architectures have been proposed to distribute heavy computation tasks such as machine learning algorithms.
In our works we refined the previous CDC scheme so as to not only to be optimal in terms of storage and communication requirement, but also in terms of computation costs. The optimal two-dimensional and three-dimensional tradeoffs can howeveronly be achieved with an exponential number of files, which might not be the case in practical implementations. To mitigate this bottleneck, we proposed suboptimal schemes that require only a significantly smaller number of files and still well-approach the optimal storage, communication, computationg tradeoff curve. Our schemes are based on computational placement-delivery arrays (CPDA), which is a subclass of PDAs as introduced for caching systems, and on a general framework on how to convert CPDAs to coded computing schemes


Coordination of Autonomous Agents
----------------------------------------------------
Important applications require devices to coordinate their actions with each other and with a random state so as to achieve a common goal. For example, in autonomous navigation systems, cars have to coordinate their speed and directions with each other and according to external obstacles. Another example are production sites of traditional and alternative energies that have to coordinate their energy productions with the consumptions of the users so as to ensure stability of the power grid.
In this project, we identified the minimum amount of common randomness or broadcast communication rate required in order that a central entity can coordinate the actions of two different terminals.
In the first 54 months of this project we have made significant progress on context- and task-oriented communication techniques for various timely and important applications. In the last 6 months of the project we will further work on improved coding techniques for wireless distributed computing systems. We also expect some more results on the joint sensing and communication problem as well as on hypothesis testing problems with multiple decision centres.
A Distributed Hypothesis Testing System
Mixed-Delay Traffic Types in Modern Communication
A Distributed Computing System with Straggling Nodes
Coordination of Autonomous Agents/Joint Sensing and Communication Situation
Age Of Information as a Relevant Information