No abstract

Material scientists are increasingly adopting the use of machine learning (ML) for making potentially important decisions, such as, discovery, development, optimization, synthesis and characterization of materials. However, despite ML's impressive performance in commercial applications, several unique challenges exist when applying ML in materials science applications. In such a context, the contributions of this work are twofold. First, we identify common pitfalls of existing ML techniques when learning from underrepresented/imbalanced material data. Specifically, we show that with imbalanced data, standard methods for assessing quality of ML models break down and lead to misleading conclusions. Furthermore, we find that the model's own confidence score cannot be trusted and model introspection methods (using simpler models) do not help as they result in loss of predictive performance (reliability-explainability trade-off). Second, to overcome these challenges, we propose a general-purpose explainable and reliable machine-learning framework. Specifically, we propose a novel pipeline that employs an ensemble of simpler models to reliably predict material properties. We also propose a transfer learning technique and show that the performance loss due to models' simplicity can be overcome by exploiting correlations among different material properties. A new evaluation metric and a trust score to better quantify the confidence in the predictions are also proposed. To improve the interpretability, we add a rationale generator component to our framework which provides both model-level and decision-level explanations. Finally, we demonstrate the versatility of

No abstract

Background: Palliative care is a patient-centred, integrated approach for improving quality of life for both patients facing life-threatening illnesses and for their families. Although there has been increased interest in palliative care for non-cancer patients, the palliative care competency of nurses who care for non-cancer patients has rarely been investigated. This study described the palliative care knowledge, attitude, confidence, and educational needs in nurses who care for patients with congestive heart failure, stroke, end-stage renal disease, and end-stage liver disease; explored the relationships between those variables; and identified factors affecting nurses' palliative care confidence. Methods: A cross-sectional, descriptive, correlational design was employed; data collection was conducted at a tertiary hospital in Seoul, Korea. Nurses who were working in general wards and intensive care units (N = 102) completed valid and reliable self-administered questionnaires. Descriptive statistics, frequencies, independent t-tests, one-way ANOVA, Pearson's correlations, and multiple regression were conducted to analyse the data. Results: Nurses' palliative care knowledge level was low (9.73 ± 2.10; range = 0-20) and their attitude toward palliative care was moderate (87.97 ± 6.93, range: 30-120). Knowledge was significantly correlated with attitude (r = .29, p = .003). Nurses were highly confident in pain and symptom management but demonstrated high educational needs for managing human and material resources to provide palliative care. Previous training in hospice, palliative, and EOL care was a significant and modifiable factor that affected nurses' confidence (std. β = 0.25, p = .010). Conclusions: To facilitate high-quality palliative care for non-cancer patients and families, nursing education programs should be developed to address nurses' knowledge level, confidence level, and educational needs. This study provides relevant information that can be utilised to develop palliative care educational programs for nurses who care for non-cancer patients.

No abstract