SANTM: Efficient Self-attention-driven Network for Text Matching

Self-attention mechanisms have recently been embraced for a broad range of text-matching applications. Self-attention model takes only one sentence as an input with no extra information, i.e., one can utilize the final hidden state or pooling. However, text-matching problems can be interpreted either in symmetrical or asymmetrical scopes. For instance, paraphrase detection is an asymmetrical task, while textual entailment classification and question-answer matching are considered asymmetrical tasks. In this article, we leverage attractive properties of self-attention mechanism and proposes an attention-based network that incorporates three key components for inter-sequence attention: global pointwise features, preceding attentive features, and contextual features while updating the rest of the components. Our model follows evaluation on two benchmark datasets cover tasks of textual entailment and question-answer matching. The proposed efficient Self-attention-driven Network for Text Matching outperforms the state of the art on the Stanford Natural Language Inference and WikiQA datasets with much fewer parameters.

Patterns of production and sustainability of cattle ranching in the state of Pará - Brazilian Amazon

The aim of this study was to analyze the production patterns present in rural properties producing cattle in the micro-regions that make up the state of Pará. Exploratory Factor Analysis (EFA) was applied to identify the patterns, and these data are used to evaluate correlation between the heterogeneity of rural properties and the environmental impact on the identified patterns. The theoretical contribution is based on discussions on global impacts of food production and environmental sustainability and the impacts of livestock production systems in Brazil and the Amazon. Survey data were taken from the 2017 Agricultural Census, available for the 144 municipalities in the state, and pooled into 22 micro-regions. Three patterns of rural properties were identified: the first related to conservation management practices and called transition management; the second highlights aspects associated with information technology and communication (ICT) and productivity called technical productive efficiency; the third indicates the importance of social organization and access to information called social participation. With these patterns, it was possible to develop the Traditional Performance Indicator (TPI), in which the micro-regions of São Félix do Xingu, Itaituba and Conceição do Araguaia were those with the highest values of this indicator, water protection practices are present in the properties, however, in precarious conditions, and conservation practices are rarely used. The correlation between heterogeneity, measured by the size of pasture area in each microregion, and the TPI is positive, strong and significant. In this transition context, public policies are essential to provide access to infrastructure, credit and good animal health and biotechnology practices

Temporal Relation Extraction in Clinical Texts

Unstructured data in electronic health records, represented by clinical texts, are a vast source of healthcare information because they describe a patient's journey, including clinical findings, procedures, and information about the continuity of care. The publication of several studies on temporal relation extraction from clinical texts during the last decade and the realization of multiple shared tasks highlight the importance of this research theme. Therefore, we propose a review of temporal relation extraction in clinical texts. We analyzed 105 articles and verified that relations between events and document creation time, a coarse temporality type, were addressed with traditional machine learning–based models with few recent initiatives to push the state-of-the-art with deep learning–based models. For temporal relations between entities (event and temporal expressions) in the document, factors such as dataset imbalance because of candidate pair generation and task complexity directly affect the system's performance. The state-of-the-art resides on attention-based models, with contextualized word representations being fine-tuned for temporal relation extraction. However, further experiments and advances in the research topic are required until real-time clinical domain applications are released. Furthermore, most of the publications mainly reside on the same dataset, hindering the need for new annotation projects that provide datasets for different medical specialties, clinical text types, and even languages.

Forage sorghum production under rainfed farming in sites with and without climate restriction for maize cultivation

Forage sorghum is a crop that can be planted in semiarid regions, due to its greater adaptability to dry climate environments, and can replace maize in these regions, which are often unsuitable for its production. Thus, the objective of the study was to evaluate the structural, morphological and nutritional characteristics of 23 sorghum hybrids forage cultivated in rainfed conditions, planted in different climate conditions, comparing the hybrids, in order to determine what produces the best in the climatic conditions of the explored region, and also to indicate whether this crop can be planted as a replacement for maize in environments not suitable for planting it. The research was conducted in climate BSh in the Municipality of Alvorada do Gurgueia, and climate Aw in the Municipality of Bom Jesus, both in the state of Piauí from 2014 to 2015. Each trial consisted of 20 experimental forage sorghum hybrids [Sorghum bicolor (L.) Moench], and three commercial hybrids. A randomized block design was used, with three replications in a factorial scheme (2 × 23). The growth characteristics determined were hybrid × climate interaction for the variables plant height, lodging and leaf/stem ratio. For the variable number of tillers, there was a significant difference only between hybrids. There was no difference between hybrids only for the lodging variable of climate Aw. The other variables showed a difference in all hybrids evaluated. There was an interaction for production of dead matter and total dry forage mass between the different environments and hybrids evaluated. For leaf production, there was an effect only for the different environments. For the chemical characteristics, there was an interaction for all variables analyzed between the different environments and hybrids evaluated. The semi-arid region of the State of Piauí, climate BSh which presents a high climatic risk for maize cultivation, proved to be favorable for forage sorghum production. The forage sorghum also presented agronomic characteristics similar to those found in semi humid climate Aw, a favorable region for maize cultivation. In addition, the tested hybrids showed good chemical characteristics, so the BSh climate has great exploratory potential for the cultivation of forage sorghum.

Readmission Prediction for Patients with Heterogeneous Medical History: A Trajectory-Based Deep Learning Approach

Hospital readmission refers to the situation where a patient is re-hospitalized with the same primary diagnosis within a specific time interval after discharge. Hospital readmission causes $26 billion preventable expenses to the U.S. health systems annually and often indicates suboptimal patient care. To alleviate those severe financial and health consequences, it is crucial to proactively predict patients’ readmission risk. Such prediction is challenging because the evolution of patients’ medical history is dynamic and complex. The state-of-the-art studies apply statistical models which use static predictors in a period, failing to consider patients’ heterogeneous medical history. Our approach – Trajectory-BAsed DEep Learning (TADEL) – is motivated to tackle the deficiencies of the existing approaches by capturing dynamic medical history. We evaluate TADEL on a five-year national Medicare claims dataset including 3.6 million patients per year over all hospitals in the United States, reaching an F1 score of 87.3% and an AUC of 88.4%. Our approach significantly outperforms all the state-of-the-art methods. Our findings suggest that health status factors and insurance coverage are important predictors for readmission. This study contributes to IS literature and analytical methodology by formulating the trajectory-based readmission prediction problem and developing a novel deep-learning-based readmission risk prediction framework. From a health IT perspective, this research delivers implementable methods to assess patients’ readmission risk and take early interventions to avoid potential negative consequences.