COPA 2024: Papers with Abstracts

Abstract. I have spent my long career developing the CSP process algebra, with much attention going to make it usable in real-world problems. In this paper I reflect on this work and try to sum up where we are in 2024. I concentrate on some relatively recent applications including the ideas behind the Coco System and how CSP can be used to support decentralised reasoning in the presence of Byzantine behaviour. I think CSP models are the ideal starting point when you want to get to grips with a challenging issue in practical concurrency.

Abstract. I consider the problem of attempting to constrain the behaviour of artificial intelligence software to prevent it from carrying out unwanted, dangerous, or malicious operations. I shall analyse a medical appliance scenario: the artificial pancreas. I shall model a diabetes patient as a system which contains a certain level of insulin and a certain level of glucose at any given point in time, and may be engaged in eating, rest, or exercise. Sugar levels will be lowered using up some insulin. Exercise will also contribute to lowering sugar levels. The artificial pancreas will periodically inject some more insulin into the patient depending on decisions made by an AI-driven control system when it is provided with monitoring data and it is intended to learn over time how best to manage the patient’s glucose levels. There is also a safety system which may decide to suspend the AI system and trigger a manual or algorithmic override should a particular threshold be breached. I shall use the CSP process algebra to represent the processes and the associated FDR proof tool to deduce properties about their behaviour.

Abstract. Quantum computers are notoriously difficult to program. This is largely because they operate predominantly on complex algebraic structures which are not usually part of the vocabulary of conventional programmers who are more used to working with integers, floating point numbers and Boolean values. Also, the atomic operations carried out by quantum gates require a level of understanding of counter-intuitive properties of quantum mechanics, such as superposition, measurement, and entanglement, which is a considerable step beyond the familiar Boolean logic which is implemented by gates in classical circuits. Therefore, there is a significant challenge for software developers who have been trained in how to create code for conventional computers in making the transition to quantum. Quantum programming is currently the preserve of mathematicians and theoretical physicists. In this paper I shall consider whether the process algebra of CSP could be useful tool in this context to make the field of Quantum Computing more accessible.

Abstract. CSP-based OCCAM modeling promises to bring great clarity to Internet of Things (IoT) designs that have been made vague by abstract tools. This requires an absoluteness in timing and data access that is more characteristic of the Transputer than of modern systems. However, it turns out some standard C and *n*x library tools can provide it, if human timing suffices.
IoT design is strongly dependent on use cases, and we had the fortune of a very definite project use case that drove the small OCCAM-modeled system described in the Reference. Key tools in the design include Unix sockets, ssh, and the kernel-based call select(). This paper will complete that system, begun in the referenced Fringe, and relate it to the general issues to which this technique can be expanded. Its occam modeling will be defended based on timing knowledge, and its hardware-software equivalence (HSE) will be proven.
This development must be read in conjunction with the Reference, which see. It de- velops the Reference to allow three or more independent racing systems. It explains the applicability of the revised approach to general IoT.

Abstract. This paper presents the Adaptive Deep Learning-Enhanced Non-Terrestrial Network (ADL-NTN), an innovative framework that combines satellites, High Altitude Platform Stations (HAPS), and Unmanned Aerial Vehicles (UAVs). By integrating Free-Space Op- tical (FSO) and Radio Frequency (RF) communications optimised for different altitudes, this architecture aims to improve connectivity in remote and disaster-affected regions. The ADL-NTN employs deep learning algorithms for dynamic power distribution and link opti- misation, significantly enhancing the network’s robustness and adaptability to environmen- tal conditions and varying demands. Simulations conducted in OMNeT++ demonstrate substantial improvements, with throughput increasing by up to 37% and latency decreas- ing by 42%, surpassing traditional NTN systems. The ADL-NTN architecture exhibits exceptional resilience, ensuring high-quality service delivery under diverse conditions. This research sets the stage for integrating future communication technologies and expanding the framework for global implementation. The ADL-NTN offers groundbreaking solutions for enhancing rural connectivity and providing rapid disaster response, significantly con- tributing to global digital inclusion

Abstract. In this paper, we delve into the creation of a mesh grid network, focusing on solving
the Santa Claus Problem in concurrent programming. It discusses how we implemented a client/server communication model and tested the system’s reliability and robustness. Our findings have significant implications for the advancement of wireless communication technologies, especially Bluetooth mesh grids, and point to the necessity of future work involving actual distributed hardware for validation.

Abstract. Neuralnetworksareomnipresentinnaturallanguageprocessing(NLP). We benchmark three popular Python frameworks (DyNet, TensorFlow, and Theano) on the standard NLP task of language modeling, and find that DyNet is significantly faster on this task. We also discuss other bottlenecks beyond performance, such as ease of use, that may impact the selection of a neural network framework.