DOI: 10.3897/rio.12.e183261

Authors: Enriko Kroon, Ricardo Borges, Rômulo de Jesus, Stefan Kuhn

Abstract: It has been shown that it is possible to train a (simple) neural network to classify nuclear magnetic resonance spectra by a substructures either being part of the chemical structure measured or not. We now explore the interpretability of such models using techniques from explainable AI, specifically Grad-CAM. We show that those techniques do not give ideal results in the context of NMR, which would be able to identify individual peaks. On the other hand, they enable a better interpretation of the results than those metrics just based on "right or wrong". We can also confirm the result from our previous work, that the trained network performs well for pure compounds, but its generalisability to mixtures is questionable, a limitation that could only be assumed in the original study.

HTML

XML

PDF

DOI: 10.3897/rio.11.e176476

Authors: Tina Heger, Stella Berger, Jonathan Jeschke, Chris Kittel, Peter Kraker, A. Makower, Daniel Mietchen, Jens Nejstgaard, Maxi Schramm

Abstract: Water is vital for life on Earth, but aquatic environments worldwide are facing critical challenges that cause severe problems for biodiversity and human well-being. These challenges include, for example, water pollution, habitat degradation, escalating water and air temperatures, salinisation of freshwaters, ocean acidification and invasive species. Since these stressors interact in complex ways, developing predictions and mitigation measures is difficult. Mesocosm experiments, offering controlled, yet realistic settings, are crucial for understanding and mitigating the impact of various stressors and their combinations on aquatic ecosystems. Mesocom facilities are key Research Infrastructures (RI), as they bridge the gap between laboratory experiments and natural systems allowing studies of highly complex environments comparable to natural ecosystems, while still offering controlled and replicated settings not available in natural systems.The AQUACOSM-RI consortium, comprising over 60 individual state-of-the-art mesocosm facilities at 28 host institutions across Europe, has therefore been instrumental in advancing aquatic environmental research across climate zones including marine, brackish and freshwater ecosystems. In addition, the EU H2020-INFRAIA projects AQUACOSM (CORDIS No. 731065) and AQUACOSM-plus (CORDIS No. 871081, www.aquacosm.eu) have developed a virtual network beyond Europe of presently > 85 host institutions with > 120 aquatic mesocosm facilities around the world, www.mesocosm.org. However, the rich, yet disconnected resources in aquatic mesocosm-based experimental research and mitigation approaches need to be combined in a modern, visible and accessible way.The project AQUANAVI (Navigating Grand Challenges and their Mitigation using Aquatic Experimental RIs) aims to enhance existing efforts by creating an interactive atlas of aquatic mesocosm facilities and related mesocosm-based experimental research. Integrating data, publications, reports and information on mesocosm facility capacities generated by the AQUACOSM consortium and other mesocosm facilities in Europe and beyond, AQUANAVI will facilitate fast discovery of resources and unused potentials of available mesocosm facilities in a modern, visible and accessible way that is presently not available. Such a multidimensional tool is expected to enable novel collaborations and a much faster setup and execution of connected and/or distributed experiments and efficient development of environmental mitigation strategies. Built upon the AQUACOSM-RIs and their encompassing data and information repository as well as scientific and technical competence, while also leveraging related infrastructures like AnaEE, EMBRC, JERICO-RI and eLTER, AQUANAVI will provide a comprehensive resource platform to more effectively explore available resources for aquatic experimental research.AQUANAVI will bridge this wealth of scientific data, expertise and mesocosm facility information through Hi Knowledge, an innovative analysis and visualisation platform that merges Wikidata, Open Knowledge Maps,and Scholia. Hi Knowledge harnesses the semantic capabilities of Wikidata to rapidly construct a FAIR and open corpus for a domain, based on a sophisticated conceptual classification system. Subsequently, Hi Knowledge incorporates visualisation components from Open Knowledge Maps and Scholia, allowing researchers to smoothly navigate information using cutting-edge visualisation techniques, artificial intelligence and knowledge synthesis methods.Open and collaborative by design, AQUANAVI’s architecture will engage a broad range of research communities. By consolidating data and information from diverse RIs, the platform will leverage and enhance the AQUACOSM and related research infrastructures, securing the reusability and interoperability of existing data collections and better exploration of existing RIs in the future. Compliant with FAIR principles and EOSC requirements, AQUANAVI will ensure the long-term sustainability and openness of its resources, enriching both the ENVRI services portfolio and the broader scientific community. In summary, AQUANAVI will empower researchers and stakeholders to implement measures to mitigate the effects of climate change and other Grand Challenges facing aquatic environments, serving as a key resource within and beyond the European research area.

HTML

XML

PDF

DOI: 10.3897/rio.2.e9232

Authors: Pekka Neittaanmäki, Timo Huttula, Juha Karvanen, Tom Frisk, Jouni Tuomisto, Antti Simola, Tero Tuovinen, Janne Ropponen

Abstract: Open data and models are becoming increasingly available, but there are not yet good methods and platforms to turn those into systematic evidence-based decision support. Unicorn will produce such an enviro­­nment based on existing theoretical and practical knowledge about decision support and models. This con­sortium possesses the necessary models, data, and skills to set up an environment and demonstrate its func­tionality and usefulness with several case studies related to the environmental issues, human health, and economy. The Unicorn environment will be built in a generic and systematic way so that it could even be­come an international standard for evidence-based decision support. Developing a technical environment or standard is not enough. Using the Unicorn environment is a large cul­­tural change for both researchers and decision makers, as the current decision support practices do not re­flect the principles of openness, criticism, or reuse. Therefore, this cultural change must be promoted by train­ing to use the environment, by informing the society about its possibilities, and solving a number of practi­cal and technical problems related to current practices in research institutes, ministries, and municipalities. We acknowledge these problems and offer solutions to them with an extensive interaction plan.

HTML

XML

PDF