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.

Boundary resource interactions in solution networks

Purpose This study aims to explore the interactions between two different and potentially complementary boundary resources in coordinating solution networks in a digital platform context: boundary spanners (those individuals who span interorganizational boundaries) and boundary interfaces (the devices that help coordinate interfirm relationships, e.g. electronic data interchanges, algorithms or chatbots). Design/methodology/approach The authors conducted a multiple case study of three firms using digital platforms to coordinate solution networks in the information communication technology and lighting facility industries. Data were collected from 30 semi-structured interviews, which are complemented by secondary data. Findings As task complexity increases, smarter digital interfaces are adopted. When the intelligence level of interfaces is low or moderate, they are only used as tools by boundary spanners or to support boundary spanners’ functions. When the intelligence level of interfaces is high or very high, boundary spanners design the interfaces and let them perform tasks autonomously. They are also sometimes employed to complement interfaces’ technological limitations and customers’ limited user ability. Research limitations/implications The industry contexts of the cases may influence the results. Qualitative case data has limited generalizability. Practical implications This study offers a practical tool for solution providers to effectively deploy boundary employees and digital technologies to offer diverse customized solutions simultaneously. Originality This study contributes to the solution business literature by putting forward a framework of boundary resource interactions in coordinating solution networks in a digital platform context. It contributes to the boundary spanning literature by revealing the shifting functions of boundary spanners and boundary interfaces.

Anterior Cruciate Ligament Reconstructed Patients Who Recovered Normal Postural Control Have Dissimilar Brain Activation Patterns Compared to Healthy Controls

Postural control, which is a fundamental functional skill, reflects integration and coordination of sensory information. Damaged anterior cruciate ligament (ACL) may alter neural activation patterns in the brain, despite patients’ surgical reconstruction (ACLR). However, it is unknown whether ACLR patients with normal postural control have persistent neural adaptation in the brain. Therefore, we explored theta (4–8 Hz) and alpha-2 (10–12 Hz) oscillation bands at the prefrontal, premotor/supplementary motor, primary motor, somatosensory, and primary visual cortices, in which electrocortical activation is highly associated with goal-directed decision-making, preparation of movement, motor output, sensory input, and visual processing, respectively, during first 3 s of a single-leg stance at two different task complexities (stable/unstable) between ACLR patients and healthy controls. We observed that ACLR patients showed similar postural control ability to healthy controls, but dissimilar neural activation patterns in the brain. To conclude, we demonstrated that ACLR patients may rely on more neural sources on movement preparation in conjunction with sensory feedback during the early single-leg stance period relative to healthy controls to maintain postural control. This may be a compensatory protective mechanism to accommodate for the altered sensory inputs from the reconstructed knee and task complexity. Our study elucidates the strategically different brain activity utilized by ACLR patients to sustain postural control.

Multiterminal Pathfinding in Practical VLSI Systems with Deep Neural Networks

A multiterminal obstacle-avoiding pathfinding approach is proposed. The approach is inspired by deep image learning. The key idea is based on training a conditional generative adversarial network (cGAN) to interpret a pathfinding task as a graphical bitmap and consequently map a pathfinding task onto a pathfinding solution represented by another bitmap. To enable the proposed cGAN pathfinding, a methodology for generating synthetic dataset is also proposed. The cGAN model is implemented in Python/Keras, trained on synthetically generated data, evaluated on practical VLSI benchmarks, and compared with state-of-the-art. Due to effective parallelization on GPU hardware, the proposed approach yields a state-of-the-art like wirelength and a better runtime and throughput for moderately complex pathfinding tasks. However, the runtime and throughput with the proposed approach remain constant with an increasing task complexity, promising orders of magnitude improvement over state-of-the-art in complex pathfinding tasks. The cGAN pathfinder can be exploited in numerous high throughput applications, such as, navigation, tracking, and routing in complex VLSI systems. The last is of particular interest to this work.

The effectiveness of task-based language teaching on the writing skill: a systematic research synthesis

Introduction. This systematic research synthesis investigated the effectiveness of Task-based language teaching interventions on L2 writing performance of intermediate level students. Objective. The main aim was to determine the effects of independent variable manipulations of task-based language teaching on different modes of writing measured holistically and by means of CALF constructs. Methodology. The integration of qualitative and quantitative data was carried out by means of a systematic literature search and retrieval of published articles from 2010 until September 2011. Substantive and methodological features of the studies were coded and compared for the identification of commonly used practices and trends within the Task-based language teaching and L2 writing research domain. Results. The results indicate 3 major types of task-based interventions: TBLT framework, task complexity manipulations and task planning conditions have prevailed as treatments. Task complexity treatments have had beneficial effects on measures of fluency and lexical complexity while strategic planning and planning time also favored fluency in L2 writing. In turn, TBLT framework lesson treatments yielded large effects measured as Cohen’s d. Conclusion. In spite of the wide variety of treatment conditions and outcome measures for different modes of L2 writing, support is given to the importance of the pre-task cycle stage management of TBLT for intermediate level learners.

Effects of third-language learning on inhibitory control with consideration of task complexity and language proficiency

Aims and Objectives: The present study explores the question of whether learning a third language (L3) in an English as a foreign language (EFL) classroom setting induces improved inhibitory control compared with that found in bilinguals, considering task complexity and language proficiency. Methodology: Thirty-six Chinese–English second language (L2) young adult learners and 121 Chinese–English–Japanese/French/Russian/German L3 young adult learners with three levels of L3 proficiency participated in the study. Simon arrow tasks were employed to measure two types of inhibitory control: response inhibition (the less complex task with univalent stimuli) and interference suppression (the more complex task with bivalent stimuli). Data and Analysis: Statistics using ANOVAs and multiple comparisons were employed to analyze the effects of L3 learning on the reaction time and accuracy for response inhibition and interference suppression, respectively. Findings: The results demonstrated that L3 learners did not outperform L2 learners in the two types of inhibitory control: response inhibition (less complex) and interference suppression (more complex). Moreover, L3 learners with a higher proficiency did not display better inhibitory control than those with a lower proficiency in response inhibition and interference suppression. However, as the L3 proficiency increased, some specific aspects of inhibitory control did improve and exhibited a nonlinear pattern. Originality: The present study extends bilingual advantage in inhibitory control to formal L3 learning, exploring whether bilingual advantage in inhibitory control also appears in L3 learners, considering task complexity and language proficiency. Significance/implications: The present study contributes to the theory of the relationship between multilingualism and inhibitory control by showing that this relationship may be more complex than it is understood currently. Learning an additional language to L2, particularly short-term learning, may not lead to an incremental advantage in overall inhibitory control. However, as learning time increases, changes may appear in specific aspects of inhibitory control, and may be a nonlinear one.

Regulatory focus and construal level theory on low-income consumer indebtedness: Evidence from an emerging market

PurposeThis study aims to further analyse the decision-making process of low-income consumer from an emerging market by verifying the influence of regulatory focus and construal level theory on indebtedness.Design/methodology/approachAn experimental study was carried out with a design 2 (regulatory focus: promotion vs prevention) × 2 (psychological distance: high vs low) between subjects, with 140 low-income consumers.FindingsOur study points out that the propensity towards indebtedness of low-income consumer is higher in a distal psychological distance. We found that promotion and prevention groups have the same propensity to indebtedness. Moreover, we highlight that low-income consumers are prone to propensity to indebtedness due to taking decisions focused on the present with an abstract mindset.Social implicationsFinancial awareness advertisements should focus on providing more concrete strategies in order to reduce decision-making complexity and provide ways to reduce competing situations that could deplete self-regulation resources. Also, public policy should organize educational programs to increase the low-income consumer's ability to deal with personal finances and reduce this task complexity. Finally, educational financial programs should also incorporate psychology professionals to teach mindfulness techniques applied to financial planning.Originality/valueThis study is the first to consider regulatory focus and construal level to explain low-income indebtedness. This paper provides a deeper analysis of the low-income consumers' decision process. Also, it supports and guides future academic and decision-making efforts.