- Book Chapter
- 10.1515/9783111354934-009
163Chapter 9 Generative adversarial network (GAN) frameworks and platforms
- Aug 09, 2025
- Ambika Paranthaman
Publications from 2021 to 2026
Showing 10 of 14 papers
163Chapter 9 Generative adversarial network (GAN) frameworks and platforms
SKAO observation execution tool: designing for concurrent, responsive observations
The SKA Observatory, currently in the construction phase, will have two of the world’s largest radio telescopes when completed in 2028. The scale of the project introduces unique challenges for the telescope software design and implementation at all levels, from user interfacing software down to the lower-level control of individual telescope elements. The Observation Execution Tool (OET) is part of the Observation Science Operations (OSO) suite of applications and is responsible for orchestrating the highest level of telescope control through the execution of telescope control scripts. One of the main challenges for the OET is creating a design that can robustly run concurrent observations on multiple subarrays while remaining responsive to the user. The Scaled Agile Framework (SAFe) development process followed by the SKA project also means the software should be allow to iterative implementation and easily accommodate new and changing requirements. This paper concentrates on the design decisions and challenges in the development of the OET, how we have solved some of the specific technical problems and details on how we remain flexible for future requirements.
Read moreDevelopment of the observatory software for the SKAO
The Observatory Science Operations (OSO) subsystem of the SKAO consists of a range of complex tools which will be used to propose, design, schedule and execute observations. Bridging the gap between the science and telescope domains is the key responsibility of OSO, requiring considerations of usability, performance, availability and accessibility, amongst others. This paper describes the state of the observatory software as we approach construction milestones, how the applications meet these requirements using a modern technology architecture, and challenges so far.
Read morePhase identification using smart meter data
Estimation of peak currents in LV feeders requires accurate data identifying customer service connections and the phases of each single-phase customer. The SMITN project applied correlation and machine-learning clustering techniques using smart meter voltage data to determine connection phases. Results show the impact of time-resolution and measurement duration on the accuracy of phase detections, either with or without substation monitoring, and highlight real-world improvements to smart meter voltage recording that would improve the performance of network monitoring.
Read moreResearching, Co-creating and Testing Innovations in Paper Health Information Systems (PHISICC) to Support Health Workers Decision-making: Protocol of a Multi-country, Transdisciplinary, Mixed-methods Research Programme in Three Sub-saharan Countries
Abstract BackgroundHealth Information Systems are crucial to provide data for decision-making and data demands are constantly growing. However, the link between data and decisions is not always rational nor linear and the management of data ends up overloading frontline health workers, who may have to compromise the health care. Despite limited evidence, there is an increasing push for the digitalisation of Health Information Systems, which faces enormous challenges, particularly in remote, rural settings in low- and middle-income countries. Paper-based tools will continue to be used and this warrants efforts to make them more responsive to local needs. Paper Health Information Systems (PHISICC) is a transdisciplinary, multi-country research initiative to create and test innovative paper-based Health Information Systems in three Sub-Saharan African countries.MethodsThe PHISICC initiative is taking place in remote, rural settings, in Côte d’Ivoire, Mozambique and Nigeria, through partnership with Ministries of Health and research institutions. We began with research syntheses to acquire the most up to date knowledge on Health Information Systems. These were coupled with field work in the three countries to understand the current design, patterns and contexts of use, and health care worker perspectives. Frontline health workers, with designers and researchers, used co-creation methods to produce the new PHISICC tools. This suite of tools is being tested in the three countries. Throughout the project, we have engaged with a wide range of stakeholders and have kept the highest scientific standards to keep it relevant to health policy in each of the three countries.DiscussionWe have deployed a comprehensive research approach to ensure the robustness and future policy uptake of the finding. Beyond the resulting paper-based tool design innovations, our process itself was innovative. Rather than emphasizing the data management compliance aspects we focused instead on frontline health workers’ decision-making; by tackling the whole scope of health care areas in Primary Health Care rather than incremental improvement to existing tools, we developed an entirely new design approach and language for a suite of tools in Primary Health Care. The initiative is being tested in remote, rural areas where the most vulnerable live.
Read moreESA Digital Twin Earth Precursor: Food Systems
<p>Part of ESA’s Digital Twin Earth Precursor projects, our project focuses on supporting ESA in the definition of the concept of a Digital Twin Earth, and establishing a solid scientific and technical basis to realise this. The project, run by CGI and in close collaboration with Oxford University Innovation, Trillium & IIASA, has a focus on developing a Food Systems Digital Twin, taking on board interdisciplinary systems through the biosphere, atmosphere, and hydrosphere systems. These in turn would allow for new interdisciplinary insights for policies dealing with climate, food production and sustainability. The project is looking at a use case with the prominent use of AI processing, challenges of model integration, ingestion of socio-economic as well as physical measurements, end-to-end chain providing decision support outputs, all with innovation at each stage, and working closely with a series of stakeholders.</p><p>The purpose of our use case is to demonstrate the value of the Digital Twin Earth concept to the scientific community, by integrating the outputs of novel algorithms. We will be using selected machine learning extreme precipitation models feeding Global Gridded Crop Models, and after a regional downscaling exercise, the integration into cropland land use and pricing. By taking these steps, the benefits include improvement in routine monitoring with regular seasonal progress, short term policy development including responses to crop shortages due to extremes, and aiding in long term policy development to apply appropriate incentives. The purpose of the architecture and integration within the preparation of the demonstration is to support the use case and draw conclusions for the roadmap. These developments will be based on stakeholder consultations and the drawing together of differing model elements.</p><p>This Digital Twin Earth is an exciting project bringing together EO experts, Earth System Scientists, industry, AI experts, modellers, ICT experts and user community. It aims to establish the initial building blocks of an ambitious initiative, and, based on the prototyping activities, to develop a scientific and technology roadmap for the future, addressing current limitations. It ties in closely to both the European Space Agency’s and European Commission’s plan to create a series of interdisciplinary Digital Twin Earths with associated boundary conditions, in order to offer services to public sector users for developing, monitoring and assessing the impact of proposed policy and legislative measures concerning the environment and climate.</p>
Read moreContemporaneous Monitoring of the Whole Dynamic Earth System from Space, Part I: System Simulation Study Using GEO and Molniya Orbits
Despite the wealth of data produced by previous and current Earth Observation platforms feeding climate models, weather forecasts, disaster monitoring services and countless other applications, the public still lacks the ability to access a live, true colour, global view of our planet, and nudge them towards a realisation of its fragility. The ideas behind commercialization of Earth photography from space has long been dominated by the analytical value of the imagery. What specific knowledge and actionable intelligence can be garnered from these evermore frequent revisits of the planet’s surface? How can I find a market for this analysis? However, what is rarely considered is what is the educational value of the imagery? As students and children become more aware of our several decades of advance in viewing our current planetary state, we should find mechanisms which serve their curiosity, helping to satisfy our children’s simple quest to explore and learn more about what they are seeing. The following study describes the reasons why current GEO and LEO observation platforms are inadequate to provide truly global RGB coverage on an update time-scale of 5-min and proposes an alternative, low-cost, GEO + Molniya 3U CubeSat constellation to perform such an application.
Read moreOrogenic gold prospectivity mapping using machine learning
As major mineral discoveries have become rarer over the last two decades, the industry has begun to turn to new technologies to assist in the exploration process. One such advancement is the applic...
Read moreTraining and assessing non-technical skills a practical guide
Operational Control and Managing Change: The Integration of Nontechnical Skills With Workplace Procedures
Summary Some of the common pitfalls and problems associated with managing operations are discussed, with suggestions about how these can be addressed. Existing literature on human error, operational decision making, and industrial psychology is combined with the authors' experiences in researching these problems and working with operations teams. The nontechnical skills required to manage operations effectively are described together with the organizational structures and work processes that best support operations. Techniques for training teams and helping them to develop the competencies to manage operations safely and to deal with uncertainty are explained. Mindful that “proper planning prevents poor performance,” engineers go to considerable lengths to create quality plans. A structured well-delivery process is followed, complete with peer reviews, risk assessments, and technical sign-offs. Deliverables include statements of requirements, basis-of-design documents, hazard identification (HAZID) sessions, drill-the-well-on-paper meetings, and a drilling program. Despite these preparations, teams make decisions during operations with sometimes costly consequences, especially when surprised by an unexpected event or when facing high levels of uncertainty. This problem is not unusual and typically occurs when a project team fails to prepare and actively manage its transition from planning to operations, and to anticipate the necessary change in the functions of the team. Planning and operations require completely different mindsets. The lack of preparation for this transition stems from a common assumption that the team that has performed the planning is more than capable of managing a dynamic operation effectively. Research has identified that two different modes are required—during the planning phase, teams produce a plan designed to solve a problem, whereas during operations the team has to manage a task requiring the ability to adapt dynamically to shifting demands of the situation. Effective operations management requires both a high degree of operational discipline and teams that are skilled in dealing with the unexpected and making decisions in time-constrained situations. These skills are not intuitive and must be taught, trained, practiced, and assessed. The protocols and decision rights must be clear, and the team must be trained so that it executes these functions in a fluent and competent manner. The oil industry lacks an international regulatory framework enjoyed by aviation. It lacks even basic standardization about how rigs are operated, wells are drilled, and operations are monitored. These are essential prerequisites for an environment within which a standardized operational discipline can be developed and implemented. The purpose of this paper is, therefore, to propose a framework for operational governance within which the development of nontechnical-skills training can be built.
Read more