-
OpenAi's GPT4 as coding assistant
Authors:
Lefteris Moussiades,
George Zografos
Abstract:
Lately, Large Language Models have been widely used in code generation. GPT4 is considered the most potent Large Language Model from Openai. In this paper, we examine GPT3.5 and GPT4 as coding assistants. More specifically, we have constructed appropriate tests to check whether the two systems can a) answer typical questions that can arise during the code development, b) produce reliable code, and…
▽ More
Lately, Large Language Models have been widely used in code generation. GPT4 is considered the most potent Large Language Model from Openai. In this paper, we examine GPT3.5 and GPT4 as coding assistants. More specifically, we have constructed appropriate tests to check whether the two systems can a) answer typical questions that can arise during the code development, b) produce reliable code, and c) contribute to code debugging. The test results are impressive. The performance of GPT4 is outstanding and signals an increase in the productivity of programmers and the reorganization of software development procedures based on these new tools.
△ Less
Submitted 22 September, 2023;
originally announced September 2023.
-
Metis: Multi-Agent Based Crisis Simulation System
Authors:
George Sidiropoulos,
Chairi Kiourt,
Lefteris Moussiades
Abstract:
With the advent of the computational technologies (Graphics Processing Units - GPUs) and Machine Learning, the research domain of crowd simulation for crisis management has flourished. Along with the new techniques and methodologies that have been proposed all those years, aiming to increase the realism of crowd simulation, several crisis simulation systems/tools have been developed, but most of t…
▽ More
With the advent of the computational technologies (Graphics Processing Units - GPUs) and Machine Learning, the research domain of crowd simulation for crisis management has flourished. Along with the new techniques and methodologies that have been proposed all those years, aiming to increase the realism of crowd simulation, several crisis simulation systems/tools have been developed, but most of them focus on special cases without providing users the ability to adapt them based on their needs. Towards these directions, in this paper, we introduce a novel multi-agent-based crisis simulation system for indoor cases. The main advantage of the system is its ease of use feature, focusing on non-expert users (users with little to no programming skills) that can exploit its capabilities a, adapt the entire environment based on their needs (Case studies) and set up building evacuation planning experiments with some of the most popular Reinforcement Learning algorithms. Simply put, the system's features focus on dynamic environment design and crisis management, interconnection with popular Reinforcement Learning libraries, agents with different characteristics (behaviors), fire propagation parameterization, realistic physics based on popular game engine, GPU-accelerated agents training and simulation end conditions. A case study exploiting a popular reinforcement learning algorithm, for training of the agents, presents the dynamics and the capabilities of the proposed systems and the paper is concluded with the highlights of the system and some future directions.
△ Less
Submitted 8 September, 2020;
originally announced September 2020.
-
Crowd simulation for crisis management: the outcomes of the last decade
Authors:
George Sidiropoulos,
Chairi Kiourt,
Lefteris Moussiades
Abstract:
The last few decades, crowd simulation for crisis management is highlighted as an important topic of interest for many scientific fields. As the continues evolution of computational resources increases, along with the capabilities of Artificial Intelligence, the demand for better and more realistic simulation has become more attractive and popular to scientists. Along those years, there have been…
▽ More
The last few decades, crowd simulation for crisis management is highlighted as an important topic of interest for many scientific fields. As the continues evolution of computational resources increases, along with the capabilities of Artificial Intelligence, the demand for better and more realistic simulation has become more attractive and popular to scientists. Along those years, there have been published hundreds of research articles and have been created numerous different systems that aim to simulate crowd behaviors, crisis cases and emergency evacuation scenarios. For better outcomes, recent research has focused on the separation of the problem of crisis management, to multiple research sub-fields (categories), such as the navigation of the simulated pedestrians, their psychology, the group dynamics etc. There have been extended research works suggesting new methods and techniques for those categories of problems. In this paper, we propose three main research categories, each one consist of several sub-categories, relying on crowd simulation for crisis management aspects and we present the outcomes of the last decade, focusing mostly on works exploiting multi-agent technologies. We analyze a number of technologies, methodologies, techniques, tools and systems introduced throughout the last years. A comparative review and discussion of the proposed categories is presented towards the identification of the most efficient aspects of the proposed categories. A general framework, towards the future crowd simulation for crisis management is presented based on the most efficient to yield the most realistic outcomes of the last decades. The paper is concluded with some highlights and open questions for future directions.
△ Less
Submitted 7 July, 2020; v1 submitted 1 June, 2020;
originally announced June 2020.
