The International Association of Oil and Gas Producers (IOGP) is a global upstream focused forum in which member companies identify and share best practices to achieve improvements in many areas including safety. In 2017, the IOGP launched an initiative called Project Safira with the aim of eliminating fatalities. One of the work streams within this project was to refresh, simplify and reduce the number of industry Life-Saving Rules to encourage industry standardization. The development of the original IOGP eight core and ten supplemental Life-Saving Rules was based on an analysis of thousands of fatal and high potential events. An industry team of subject matter experts, HSE, and operations professionals formed as a task force. This group conducted a comprehensive analysis of the latest 10 years of fatality data and developed an updated set incorporating the latest thinking on human performance and lessons learned from member companies’ experiences in implementation. From 2008 to 2017, 376 workers lost their lives in incidents that may have been prevented by following one of the new nine IOGP Life-Saving Rules. With the benefit of having an additional seven years of data and feedback from member companies on their adoption of the previous set of Life-Saving Rules, the IOGP has streamlined the original 18 rules down to nine, while retaining the level of applicability in fatality prevention. The Life-Saving Rules are not intended to replace company HSE management systems, but are rather aimed at complementing existing organizational processes and procedures. The rules provide simple actions in the form of ‘I statements’ which can provide a final barrier that individuals have control over, and by their own actions can prevent fatalities. Having the largest database of safety performance and fatality data in the upstream oil and gas industry, the IOGP has the ability to analyze trends and build on learning from fatal incidents on an industry-wide basis.
Na atualidade, a conexão entre as instituições de memória e seus usuários passa necessariamente pelo acesso à internet, o que representa a evolução dessas instituições no que tange à inovação na oferta de serviços a seus usuários. Este ensaio verifica a presença e a disponibilização de serviços de informação no ambiente web para bibliotecas, museus e arquivos, dialogando com a revisão de literatura sobre tecnologias e formalismos para representação da informação. Apoia-se na interoperabilidade entre as informações disponíveis para viabilizar a melhoria dos serviços, bem como para agregar valor e dar maior visibilidade às instituições. Propõe representar conexões entre as informações disponibilizadas por essas instituições, usando formalismos e vocabulários específicos da web semântica. Esta narrativa é fruto da discussão desenvolvida na disciplina “Bibliotecas na web, dados ligados e web semântica”, no âmbito do mestrado profissional em Biblioteconomia da Universidade Federal do Estado do Rio de Janeiro, no primeiro semestre de 2016.Palavras-chave: Biblioteca Museu. Arquivo. Representação da informação. Web semântica. Dados ligados.Link: http://memoriaeinformacao.casaruibarbosa.gov.br/index.php/fcrb/article/view/17/17
Machine Learning has been facing significant challenges over the last years, much of which stem from the new characteristics of Machine Learning problems, such as learning from streaming data or incorporating Human feedback into existing datasets and models. In these dynamic scenarios, data change over time and models must adapt. However, new data do not necessarily mean new patterns. The main goal of this paper is to devise a method to predict a model's performance metrics before it is trained, in order to decide whether it is worth it to train it or not. That is, will the model hold significantly better results than the current one? To address this issue we propose the use of metalearning. Specifically, we evaluate two different meta-models, one built for a specific Machine Learning problem, and another built based on many different problems, meant to be a generic meta-model, applicable to virtually any problem. In this paper we focus only on the prediction of the rmse. Results show that it is possible to accurately predict the rmse of future models, event in streaming scenarios. Moreover, results also show that it is possible to reduce the need for re-training models between 60% to 98%, depending on the problem and on the threshold used.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.