Detecting ethnicity-targeted hate speech in Russian social media texts

Ethnicity-targeted hate speech has been widely shown to influence on-the-ground inter-ethnic conflict and violence, especially in such multi-ethnic societies as Russia. Therefore, ethnicity-targeted hate speech detection in user texts is becoming an important task. However, it faces a number of unresolved problems: difficulties of reliable mark-up, informal and indirect ways of expressing negativity in user texts (such as irony, false generalization and attribution of unfavored actions to targeted groups), users’ inclination to express opposite attitudes to different ethnic groups in the same text and, finally, lack of research on languages other than English. In this work we address several of these problems in the task of ethnicity-targeted hate speech detection in Russian-language social media texts. This approach allows us to differentiate between attitudes towards different ethnic groups mentioned in the same text – a task that has never been addressed before. We use a dataset of over 2,6M user messages mentioning ethnic groups to construct a representative sample of 12K instances (ethnic group, text) that are further thoroughly annotated via a special procedure. In contrast to many previous collections that usually comprise extreme cases of toxic speech, representativity of our sample secures a realistic and, therefore, much higher proportion of subtle negativity which additionally complicates its automatic detection. We then experiment with four types of machine learning models, from traditional classifiers such as SVM to deep learning approaches, notably the recently introduced BERT architecture, and interpret their predictions in terms of various linguistic phenomena. In addition to hate speech detection with a text-level two-class approach (hate, no hate), we also justify and implement a unique instance-based three-class approach (positive, neutral, negative attitude, the latter implying hate speech). Our best results are achieved by using fine-tuned and pre-trained RuBERT combined with linguistic features, with F1-hate=0.760, F1-macro=0.833 on the text-level two-class problem comparable to previous studies, and F1-hate=0.813, F1-macro=0.824 on our unique instance-based three-class hate speech detection task. Finally, we perform error analysis, and it reveals that further improvement could be achieved by accounting for complex and creative language issues more accurately, i.e., by detecting irony and unconventional forms of obscene lexicon.     

New directions in the arts and sport? Critiquing national strategies

Abstract

Taking as its starting point the Fields of Vision initiative’s interest in promoting the potential benefits of bringing sports and arts closer together, this paper reviews how national (English) policy addresses that challenge. Four key strategic documents (the Government’s Sport Strategy and its Culture White Paper as well as the strategies of Arts Council England and Sport England) are examined. That is supplemented by the views of significant individuals from this interface, including the research network funded by the Arts and Humanities Research Council (AHRC). Noting the similar social remit ascribed to sport and the arts by the government, shortcomings in the current strategies are identified as barriers to integration. ‘Play’ and ‘movement’ are briefly discussed as integrating concepts alongside our assessment of the potential of the arts/sport nexus, in areas including aesthetic innovation, promoting health and wellbeing, and encouraging wider participation and engagement. Having challenged existing national policies the paper suggests possible future directions.     

Irony detection via sentiment-based transfer learning

Irony as a literary technique is widely used in online texts such as Twitter posts. Accurate irony detection is crucial for tasks such as effective sentiment analysis. A text’s ironic intent is defined by its context incongruity. For example in the phrase “I love being ignored”, the irony is defined by the incongruity between the positive word “love” and the negative context of “being ignored”. Existing studies mostly formulate irony detection as a standard supervised learning text categorization task, relying on explicit expressions for detecting context incongruity. In this paper we formulate irony detection instead as a transfer learning task where supervised learning on irony labeled text is enriched with knowledge transferred from external sentiment analysis resources. Importantly, we focus on identifying the hidden, implicit incongruity without relying on explicit incongruity expressions, as in “I like to think of myself as a broken down Justin Bieber – my philosophy professor.” We propose three transfer learning-based approaches to using sentiment knowledge to improve the attention mechanism of recurrent neural models for capturing hidden patterns for incongruity. Our main findings are: (1) Using sentiment knowledge from external resources is a very effective approach to improving irony detection; (2) For detecting implicit incongruity, transferring deep sentiment features seems to be the most effective way. Experiments show that our proposed models outperform state-of-the-art neural models for irony detection.     

Universities’ attractiveness to students: The Darwinism effect

Due to significant government cuts to Higher Education funding in Southern European systems, their already underfunded universities were forced to increasingly compete for students as sources of additional revenue. Concurrently, families and students that continued to afford participation in Higher Education became more selective when choosing a university, realising the riskier investment that Higher Education participation had become. Through a competing destinations model and relying on all Italian private and public universities, this study finds that the competition forces characterising universities’ attractiveness over the last decade have changed since the financial crisis of 2008. In a context of lower demand for Higher Education, the competition for students grew and universities in close proximity were better prepared to face the new challenges, leading to the growth of Higher Education clusters.     

A Graphical Method for Assessing the Identification of Linear Structural Equation Models

A graphical method is presented for assessing the state of identifiability of the parameters in a linear structural equation model based on the associated directed graph. We do not restrict attention to recursive models. In the recent literature, methods based on graphical models have been presented as a useful tool for assessing the state of identifiability of the parameters of the model. This article proposes the graphical counterpart of the rank condition of the matrix of structural coefficients, which allows for checking of the identifiability through a simple graphical rule. This approach can be used to develop algorithms.     

Crossing the Border from Science Student to Science Teacher: Preservice Teachers’ Views and Experiences Learning to Teach Inquiry

Preservice science teachers face numerous challenges in understanding and teaching science as inquiry. Over the course of their teacher education program, they are expected to move from veteran science students with little experience learning their discipline through inquiry instruction to beginning science teachers adept at implementing inquiry in their own classrooms. In this study, we used Aikenhead’s (Sci Educ 81: 217–238, 1997, Science Educ 85:180–188, 2001) notion of border crossing to describe this transition preservice teachers must make from science student to science teacher. We examined what one cohort of eight preservice secondary science teachers said, did, and wrote as they both conducted a two-part inquiry investigation and designed an inquiry lesson plan. We conducted two types of qualitative analyses. One, we drew from Costa (Sci Educ 79: 313–333, 1995) to group our preservice teacher participants into one of four types of potential science teachers. Two, we identified successes and struggles in preservice teachers’ attempts to negotiate the cultural border between veteran student and beginning teacher. In our implications, we argue that preservice teachers could benefit from explicit opportunities to navigate the border between learning and teaching science; such opportunities could deepen their conceptions of inquiry beyond those exclusively fashioned as either student or teacher.     

Science,Society and Trends in Science Education‐‐some Thoughts on the Development of Integrated Science in the Educational Systems of the World

Summaries

English

Before constructing a new curriculum for any level in integrated‐science education, it is very important to establish a basic philosophy of science teaching which reflects a symbiotic relationship between education and the socio‐economic development in a particular country of the world.

The curriculum structure of integrated‐science education should broaden the basis of curricular decision and should be designated to follow the expectations of the society in providing professional intellectual training so that education can make a significant contribution to the socio‐economic growth of the country. This basic assumption which established a link between cultural and socio‐economic growth, involves a diversification of the cultural trends of science education which must become appropriate to the economic and historical context of each country in a fast changing world of work.

Therefore, by critically removing outside influences in the transmission of scientific knowledge, there should be less uniformity in the context and methods of teaching, and considerable restructuring of education for each country should occur. Thus, considering the historical and geographical differences, a new order in international division of labour would come into being.

In such a hypothesis, a research‐oriented integrated curriculum in science education can play a central role (certainly a better one than in the traditional, discipline‐oriented scheme of science teaching) in obtaining an adequate interaction between scientific education and social problems; this is required for modern cultural and economic development of the world.     

Light-driven directional ion transport for enhanced osmotic energy harvesting

Light-driven ion (proton) transport is a crucial process both for photosynthesis of green plants and solar energy harvesting of some archaea. Here, we describe use of a TiO₂/C₃N₄ semiconductor heterojunction nanotube membrane to realize similar light-driven directional ion transport performance to that of biological systems. This heterojunction system can be fabricated by two simple deposition steps. Under unilateral illumination, the TiO₂/C₃N₄ heterojunction nanotube membrane can generate a photocurrent of about 9 μA/cm², corresponding to a pumping stream of ∼5500 ions per second per nanotube. By changing the position of TiO₂ and C₃N₄, a reverse equivalent ionic current can also be realized. Directional transport of photogenerated electrons and holes results in a transmembrane potential, which is the basis of the light-driven ion transport phenomenon. As a proof of concept, we also show that this system can be used for enhanced osmotic energy generation. The artificial light-driven ion transport system proposed here offers a further step forward on the roadmap for development of ionic photoelectric conversion and integration into other applications, for example water desalination.     

Reconfiguring confined magnetic colloids with tunable fluid transport behavior

Collective dynamics of confined colloids are crucial in diverse scenarios such as self-assembly and phase behavior in materials science, microrobot swarms for drug delivery and microfluidic control. Yet, fine-tuning the dynamics of colloids in microscale confined spaces is still a formidable task due to the complexity of the dynamics of colloidal suspension and to the lack of methodology to probe colloids in confinement. Here, we show that the collective dynamics of confined magnetic colloids can be finely tuned by external magnetic fields. In particular, the mechanical properties of the confined colloidal suspension can be probed in real time and this strategy can be also used to tune microscale fluid transport. Our experimental and theoretical investigations reveal that the collective configuration characterized by the colloidal entropy is controlled by the colloidal concentration, confining ratio and external field strength and direction. Indeed, our results show that mechanical properties of the colloidal suspension as well as the transport of the solvent in microfluidic devices can be controlled upon tuning the entropy of the colloidal suspension. Our approach opens new avenues for the design and application of drug delivery, microfluidic logic, dynamic fluid control, chemical reaction and beyond.