More Like this

1. Designing a Platform for Ethical Citizen Science: A Case Study of CitSci.org

   https://doi.org/http://doi.org/10.5334/cstp.227  

   Lynn, S.J. ; Kaplan, N. ; Newman, S. ; Scarpino, R. ; Newman, G. ( March 2019 , Citizen science)

   Involving the public in scientific discovery offers opportunities for engagement, learning, participation, and action. Since its launch in 2007, the CitSci.org platform has supported hundreds of community-driven citizen science projects involving thousands of participants who have generated close to a million scientific measurements around the world. Members using CitSci.org follow their curiosities and concerns to develop, lead, or simply participate in research projects. While professional scientists are trained to make ethical determinations related to the collection of, access to, and use of information, citizen scientists and practitioners may be less aware of such issues and more likely to become involved in ethical dilemmas. In this era of big and open data, where data sharing is encouraged and open science is promoted, privacy and openness considerations can often be overlooked. Platforms that support the collection, use, and sharing of data and personal information need to consider their responsibility to protect the rights to and ownership of data, the provision of protection options for data and members, and at the same time provide options for openness. This requires critically considering both intended and unintended consequences of the use of platforms, data, and volunteer information. Here, we use our journey developing CitSci.org tomore »argue that incorporating customization into platforms through flexible design options for project managers shifts the decision-making from top-down to bottom-up and allows project design to be more responsive to goals. To protect both people and data, we developed—and continue to improve—options that support various levels of “open” and “closed” access permissions for data and membership participation. These options support diverse governance styles that are responsive to data uses, traditional and indigenous knowledge sensitivities, intellectual property rights, personally identifiable information concerns, volunteer preferences, and sensitive data protections. We present a typology for citizen science openness choices, their ethical considerations, and strategies that we are actively putting into practice to expand privacy options and governance models based on the unique needs of individual projects using our platform.« less
2. Cultural imaginaries or incommensurable ontologies? Relationality and sovereignty as worldviews in socio-technological system transitions

   Schelly, C ; Gagnon, VS ; Arola, K ; Fiss, A ; Schaefer, M ; Halvorsen, KE ( October 2021 , Energy research social science)

   Scholars bridging the fields of science and technology studies (STS) and energy research in social sciences (ERSS) offer a rich and integrated conceptualization of how energy systems are imbued in social systems, including cultures, social structures, institutions, and social relations of power. Yet as fields of study, STS and ERSS are dominated by approaches to understanding nature, culture, and relationships among them with origins in western European Enlightenment thinking. In this article, we argue that the language of “imaginaries” provides an understanding of culturally organized normative commitments but may obscure attention to what are actually diverse and sometimes incommensurable yet legitimate plural ontologies. Tribal Nations, Indigenous communities, and other non-Western worldviews are not simply imagined; they offer different teachings regarding the relational and embedded realities governing relations among human and more-than-human beings across time and space. The field of STS has a rich history of exploring ontological controversies and provides insight into understanding diverse and competing perspectives in science and technology, yet without articulating the connection between this conceptual terrain and the lived realities of socio-technological system entrenchment or change. ERSS recognizes participation, energy system democratization, and even co-production as components of a just energy transition, while most typically thinkingmore »about participation as a methodology or research approach rather than as requiring consideration and even wholesale reconceptualization of ontological foundations. To advance convergent, transdisciplinary social science research in socio technological transitions requires grappling with plural ontologies regarding the reality of relations in the world. Here, we explore diverse ontologies shaping the realities of energy systems through the lens of Tribal Nations in the Great Lakes region in the United States. Ontologies that recognize reciprocal relationships among human and more-than-human beings as well as the sovereignty of these beings and their collective kinships suggest fundamentally different priorities for energy systems transitions. Moving beyond the language of imagination to recognize that cultures can involve diverse and sometimes incommensurable pluralistic ontologies is essential for developing inclusive and just frameworks for socio-technological system transitions.« less
3. Neural Polysynthetic Language Modelling

   Schwartz, Lane ; Tyers, Francis ; Levin, Lori ; Kirov, Christo ; Littell, Patrick ; Lo, Chi-kiu ; Prud'hommeaux, Emily ; Park, Hyunji Hayley ; Steimel, Kenneth ; Knowles, Rebecca ; et al ( May 2020 , Final Report of the Frederick Jelinek Memorial Summer Workshop)

   Many techniques in modern computational linguistics and natural language processing (NLP) make the assumption that approaches that work well on English and other widely used European (and sometimes Asian) languages are “language agnostic” – that is that they will also work across the typologically diverse languages of the world. In high-resource languages, especially those that are analytic rather than synthetic, a common approach is to treat morphologically-distinct variants of a common root (such as dog and dogs) as completely independent word types. Doing so relies on two main assumptions: that there exist a limited number of morphological inflections for any given root, and that most or all of those variants will appear in a large enough corpus (conditioned on assumptions about domain, etc.) so that the model can adequately learn statistics about each variant. Approaches like stemming, lemmatization, morphological analysis, subword segmentation, or other normalization techniques are frequently used when either of those assumptions are likely to be violated, particularly in the case of synthetic languages like Czech and Russian that have more inflectional morphology than English. Within the NLP literature, agglutinative languages like Finnish and Turkish are commonly held up as extreme examples of morphological complexity that challenge commonmore »modelling assumptions. Yet, when considering all of the world’s languages, Finnish and Turkish are closer to the average case in terms of synthesis. When we consider polysynthetic languages (those at the extreme of morphological complexity), even approaches like stemming, lemmatization, or subword modelling may not suffice. These languages have very high numbers of hapax legomena (words appearing only once in a corpus), underscoring the need for appropriate morphological handling of words, without which there is no hope for a model to capture enough statistical information about those words. Moreover, many of these languages have only very small text corpora, substantially magnifying these challenges. To this end, we examine the current state-of-the-art in language modelling, machine translation, and predictive text completion in the context of four polysynthetic languages: Guaraní, St. Lawrence Island Yupik, Central Alaskan Yup’ik, and Inuktitut. We have a particular focus on Inuit-Yupik, a highly challenging family of endangered polysynthetic languages that ranges geographically from Greenland through northern Canada and Alaska to far eastern Russia. The languages in this family are extraordinarily challenging from a computational perspective, with pervasive use of derivational morphemes in addition to rich sets of inflectional suffixes and phonological challenges at morpheme boundaries. Finally, we propose a novel framework for language modelling that combines knowledge representations from finite-state morphological analyzers with Tensor Product Representations (Smolensky, 1990) in order to enable successful neural language models capable of handling the full linguistic variety of typologically variant languages.« less
4. Neural Polysynthetic Language Modelling

   Schwartz, Lane ; Tyers, Francis ; Levin, Lori ; Kirov, Christo ; Littell, Patrick ; Lo, Chi-kiu ; Prud'hommeaux, Emily ; Park, Hyunji Hayley ; Steimel, Kenneth ; Knowles, Rebecca ; et al ( May 2020 , ArXivorg)

   Many techniques in modern computational linguistics and natural language processing (NLP) make the assumption that approaches that work well on English and other widely used European (and sometimes Asian) languages are “language agnostic” – that is that they will also work across the typologically diverse languages of the world. In high-resource languages, especially those that are analytic rather than synthetic, a common approach is to treat morphologically-distinct variants of a common root (such as dog and dogs) as completely independent word types. Doing so relies on two main assumptions: that there exist a limited number of morphological inflections for any given root, and that most or all of those variants will appear in a large enough corpus (conditioned on assumptions about domain, etc.) so that the model can adequately learn statistics about each variant. Approaches like stemming, lemmatization, morphological analysis, subword segmentation, or other normalization techniques are frequently used when either of those assumptions are likely to be violated, particularly in the case of synthetic languages like Czech and Russian that have more inflectional morphology than English. Within the NLP literature, agglutinative languages like Finnish and Turkish are commonly held up as extreme examples of morphological complexity that challenge commonmore »modelling assumptions. Yet, when considering all of the world’s languages, Finnish and Turkish are closer to the average case in terms of synthesis. When we consider polysynthetic languages (those at the extreme of morphological complexity), even approaches like stemming, lemmatization, or subword modelling may not suffice. These languages have very high numbers of hapax legomena (words appearing only once in a corpus), underscoring the need for appropriate morphological handling of words, without which there is no hope for a model to capture enough statistical information about those words. Moreover, many of these languages have only very small text corpora, substantially magnifying these challenges. To this end, we examine the current state-of-the-art in language modelling, machine translation, and predictive text completion in the context of four polysynthetic languages: Guaraní, St. Lawrence Island Yupik, Central Alaskan Yup’ik, and Inuktitut. We have a particular focus on Inuit-Yupik, a highly challenging family of endangered polysynthetic languages that ranges geographically from Greenland through northern Canada and Alaska to far eastern Russia. The languages in this family are extraordinarily challenging from a computational perspective, with pervasive use of derivational morphemes in addition to rich sets of inflectional suffixes and phonological challenges at morpheme boundaries. Finally, we propose a novel framework for language modelling that combines knowledge representations from finite-state morphological analyzers with Tensor Product Representations (Smolensky, 1990) in order to enable successful neural language models capable of handling the full linguistic variety of typologically variant languages.« less
5. Means, motive, and opportunity: A method for understanding stakeholder agency within food–energy–water systems

   https://doi.org/10.1525/elementa.2021.00066  

   Bird, K. I. ; Nichols, V. A. ; Garay, E. S. ; Nowatzke, M. ; Essary, C. ; Post, K. K. ; Deviney, A. V. ; Welles, J. S. ; Alao, M. O. ; Dorbu, F. E. ; et al ( March 2022 , Elementa)

   Wicked problems are inherent in food–energy–water systems (FEWS) due to the complexity and interconnectedness of these systems, and addressing these challenges necessitates the involvement of the diverse stakeholders in FEWS. However, successful stakeholder engagement requires a strong understanding of the relationships between stakeholders and the specific wicked problem. To better account for these relationships, we adapted a means, motive, and opportunity (MMO) framework to develop a method of stakeholder analysis that evaluates the agency of stakeholders related to a wicked problem in FEWS. This method involves two key components: (1) identification of a challenge at the FEWS nexus and (2) evaluation of stakeholder agency related to the challenge using the dimensions of MMO. This approach provides a method for understanding the characteristics of stakeholders in FEWS and provides information that could be used to inform stakeholder engagement in efforts to address wicked problems at the FEWS nexus. In this article, we present the stakeholder analysis method and describe an example application of the MMO method by examining stakeholder agency related to the adoption of improved swine waste management technology in North Carolina, USA.