scholarly journals Effectiveness, Explainability and Reliability of Machine Meta-Learning Methods for Predicting Mortality in Patients with COVID-19: Results of the Brazilian COVID-19 Registry

2022 ◽  
Author(s):  
Polianna Delfino-Pereira ◽  
Cláudio Moisés Valiense De Andrade ◽  
Virginia Mara Reis Gomes ◽  
Maria Clara Pontello Barbosa Lima ◽  
Maira Viana Rego Souza-Silva ◽  
...  
Keyword(s):  
Statistical Methods ◽  
State Of The Art ◽  
Prognostic Scores ◽  
Early Assessment ◽  
Meta Learning ◽  
Health Related ◽  
Previous State ◽  
Using Data ◽  
Selection Operator ◽  
Evaluation Metric

Abstract The majority prognostic scores proposed for early assessment of coronavirus disease 19 (COVID-19) patients are bounded by methodological flaws. Our group recently developed a new risk score - ABC2SPH - using traditional statistical methods (least absolute shrinkage and selection operator logistic regression - LASSO). In this article, we provide a thorough comparative study between modern machine learning (ML) methods and state-of-the-art statistical methods, represented by ABC2SPH, in the task of predicting in-hospital mortality in COVID-19 patients using data upon hospital admission. We overcome methodological and technological issues found in previous similar studies, while exploring a large sample (5,032 patients). Additionally, we take advantage of a large and diverse set of methods and investigate the effectiveness of applying meta-learning, more specifically Stacking, in order to combine the methods' strengths and overcome their limitations. In our experiments, our Stacking solutions improved over previous state-of-the-art by more than 26% in predicting death, achieving 87.1% of AUROC and MacroF1 of 73.9%. We also investigated issues related to the interpretability and reliability of the predictions produced by the most effective ML methods. Finally, we discuss the adequacy of AUROC as an evaluation metric for highly imbalanced and skewed datasets commonly found in health-related problems.

scholarly journals Goal Recognition for Deceptive Human Agents through Planning and Gaze

2021 ◽  
Vol 71 ◽  
pp. 697-732
Author(s):  
Thao Le ◽  
Ronal Singh ◽  
Tim Miller
Keyword(s):  
State Of The Art ◽  
Eye Gaze ◽  
Competitive Game ◽  
Gaze Patterns ◽  
Gaze Behaviour ◽  
Previous State ◽  
Goal Recognition ◽  
Recognition Systems ◽  
Using Data ◽  
Insight Into

Eye gaze has the potential to provide insight into the minds of individuals, and this idea has been used in prior research to improve human goal recognition by combining human's actions and gaze. However, most existing research assumes that people are rational and honest. In adversarial scenarios, people may deliberately alter their actions and gaze, which presents a challenge to goal recognition systems. In this paper, we present new models for goal recognition under deception using a combination of gaze behaviour and observed movements of the agent. These models aim to detect when a person is deceiving by analysing their gaze patterns and use this information to adjust the goal recognition. We evaluated our models in two human-subject studies: (1) using data collected from 30 individuals playing a navigation game inspired by an existing deception study and (2) using data collected from 40 individuals playing a competitive game (Ticket To Ride). We found that one of our models (Modulated Deception Gaze+Ontic) offers promising results compared to the previous state-of-the-art model in both studies. Our work complements existing adversarial goal recognition systems by equipping these systems with the ability to tackle ambiguous gaze behaviours.

scholarly journals Fair algorithms for selecting citizens’ assemblies

Nature ◽  
2021 ◽  
Author(s):  
Bailey Flanigan ◽  
Paul Gölz ◽  
Anupam Gupta ◽  
Brett Hennig ◽  
Ariel D. Procaccia
Keyword(s):  
Random Process ◽  
State Of The Art ◽  
Fair Division ◽  
High Impact ◽  
Participation Rates ◽  
The World ◽  
Previous State ◽  
Using Data ◽  
Selection Probabilities ◽  
Selection Algorithms

AbstractGlobally, there has been a recent surge in ‘citizens’ assemblies’1, which are a form of civic participation in which a panel of randomly selected constituents contributes to questions of policy. The random process for selecting this panel should satisfy two properties. First, it must produce a panel that is representative of the population. Second, in the spirit of democratic equality, individuals would ideally be selected to serve on this panel with equal probability2,3. However, in practice these desiderata are in tension owing to differential participation rates across subpopulations4,5. Here we apply ideas from fair division to develop selection algorithms that satisfy the two desiderata simultaneously to the greatest possible extent: our selection algorithms choose representative panels while selecting individuals with probabilities as close to equal as mathematically possible, for many metrics of ‘closeness to equality’. Our implementation of one such algorithm has already been used to select more than 40 citizens’ assemblies around the world. As we demonstrate using data from ten citizens’ assemblies, adopting our algorithm over a benchmark representing the previous state of the art leads to substantially fairer selection probabilities. By contributing a fairer, more principled and deployable algorithm, our work puts the practice of sortition on firmer foundations. Moreover, our work establishes citizens’ assemblies as a domain in which insights from the field of fair division can lead to high-impact applications.

Multi-Resolution Autoregressive Graph-to-Graph Translation for Molecules

2019 ◽  
Author(s):  
Wengong Jin ◽  
Regina Barzilay ◽  
Tommi S Jaakkola
Keyword(s):  
Drug Discovery ◽  
State Of The Art ◽  
Molecular Graph ◽  
Biochemical Properties ◽  
Large Margin ◽  
Previous State ◽  
Translation Methods ◽  
Atom Level ◽  
Precursor Molecules ◽  
Prior State

The problem of accelerating drug discovery relies heavily on automatic tools to optimize precursor molecules to afford them with better biochemical properties. Our work in this paper substantially extends prior state-of-the-art on graph-to-graph translation methods for molecular optimization. In particular, we realize coherent multi-resolution representations by interweaving trees over substructures with the atom-level encoding of the original molecular graph. Moreover, our graph decoder is fully autoregressive, and interleaves each step of adding a new substructure with the process of resolving its connectivity to the emerging molecule. We evaluate our model on multiple molecular optimization tasks and show that our model outperforms previous state-of-the-art baselines by a large margin.

scholarly journals Using spatial-temporal ensembles of convolutional neural networks for lumen segmentation in ureteroscopy

2021 ◽  
Author(s):  
Jorge F. Lazo ◽  
Aldo Marzullo ◽  
Sara Moccia ◽  
Michele Catellani ◽  
Benoit Rosa ◽  
...  
Keyword(s):  
Neural Networks ◽  
Convolutional Neural Networks ◽  
State Of The Art ◽  
Automatic Segmentation ◽  
Temporal Information ◽  
Invasive Technique ◽  
Dice Similarity Coefficient ◽  
Specular Reflections ◽  
Lumen Segmentation ◽  
Previous State

Abstract Purpose Ureteroscopy is an efficient endoscopic minimally invasive technique for the diagnosis and treatment of upper tract urothelial carcinoma. During ureteroscopy, the automatic segmentation of the hollow lumen is of primary importance, since it indicates the path that the endoscope should follow. In order to obtain an accurate segmentation of the hollow lumen, this paper presents an automatic method based on convolutional neural networks (CNNs). Methods The proposed method is based on an ensemble of 4 parallel CNNs to simultaneously process single and multi-frame information. Of these, two architectures are taken as core-models, namely U-Net based in residual blocks ($$m_1$$ m 1 ) and Mask-RCNN ($$m_2$$ m 2 ), which are fed with single still-frames I(t). The other two models ($$M_1$$ M 1 , $$M_2$$ M 2 ) are modifications of the former ones consisting on the addition of a stage which makes use of 3D convolutions to process temporal information. $$M_1$$ M 1 , $$M_2$$ M 2 are fed with triplets of frames ($$I(t-1)$$ I ( t - 1 ) , I(t), $$I(t+1)$$ I ( t + 1 ) ) to produce the segmentation for I(t). Results The proposed method was evaluated using a custom dataset of 11 videos (2673 frames) which were collected and manually annotated from 6 patients. We obtain a Dice similarity coefficient of 0.80, outperforming previous state-of-the-art methods. Conclusion The obtained results show that spatial-temporal information can be effectively exploited by the ensemble model to improve hollow lumen segmentation in ureteroscopic images. The method is effective also in the presence of poor visibility, occasional bleeding, or specular reflections.

The Disciplinary Differences in the Characteristics and Effects of Non-Tenure-Track Faculty

2021 ◽  
pp. 016237372110304
Author(s):  
Di Xu ◽  
Florence Xiaotao Ran
Keyword(s):  
Tenure Track ◽  
Subsequent Performance ◽  
State College ◽  
College System ◽  
Tenure Track Faculty ◽  
Engineering Mathematics ◽  
Health Related ◽  
Student Sorting ◽  
Using Data

Using data with detailed instructor employment information from a state college system, this study examines disciplinary variations in the characteristics and effects of non-tenure-track faculty hired through temporary and long-term employment. We identify substantial differences in demographic and employment characteristics between the two types of non-tenure-line faculty, where the differences are most pronounced in science, technology, engineering, mathematics, and health-related fields (STEM) at 4-year colleges. Using an instrumental variables strategy to address student sorting, our analyses indicate that taking introductory courses with temporary adjuncts reduces subsequent interest, and the effects are particularly large in STEM fields at 4-year colleges. Long-term non-tenure faculty are generally comparable with tenure-track faculty in student subsequent interest, but tenure-track faculty are associated with better subsequent performance in a handful of fields.

Artificial Intelligence, Social Media and Depression. ‘Patient’ Autonomy Revisited

2021 ◽  
Vol 66 (Special Issue) ◽  
pp. 133-133
Author(s):  
Regina Mueller ◽  
◽  
Sebastian Laacke ◽  
Georg Schomerus ◽  
Sabine Salloch ◽  
...  
Keyword(s):  
Artificial Intelligence ◽  
Social Media ◽  
Mental Disorders ◽  
Patient Autonomy ◽  
Ethical Challenges ◽  
Traditional Concept ◽  
Ethical Implications ◽  
Health Related ◽  
Twitter Users ◽  
Using Data

"Artificial Intelligence (AI) systems are increasingly being developed and various applications are already used in medical practice. This development promises improvements in prediction, diagnostics and treatment decisions. As one example, in the field of psychiatry, AI systems can already successfully detect markers of mental disorders such as depression. By using data from social media (e.g. Instagram or Twitter), users who are at risk of mental disorders can be identified. This potential of AI-based depression detectors (AIDD) opens chances, such as quick and inexpensive diagnoses, but also leads to ethical challenges especially regarding users’ autonomy. The focus of the presentation is on autonomy-related ethical implications of AI systems using social media data to identify users with a high risk of suffering from depression. First, technical examples and potential usage scenarios of AIDD are introduced. Second, it is demonstrated that the traditional concept of patient autonomy according to Beauchamp and Childress does not fully account for the ethical implications associated with AIDD. Third, an extended concept of “Health-Related Digital Autonomy” (HRDA) is presented. Conceptual aspects and normative criteria of HRDA are discussed. As a result, HRDA covers the elusive area between social media users and patients. "

Bidirectional Association Between Depression and Hearing Loss: Evidence From the China Health and Retirement Longitudinal Study

2021 ◽  
pp. 073346482110423
Author(s):  
Chao Wu
Keyword(s):  
Hearing Loss ◽  
Longitudinal Study ◽  
Cognitive Decline ◽  
Age Related ◽  
Bidirectional Association ◽  
Health Related ◽  
Clinically Significant ◽  
Using Data ◽  
Bidirectional Associations ◽  
Age Related Hearing Loss

The relationship between depression and age-related hearing loss (ARHL) is not fully understood. This study tested the bidirectional associations between clinically significant depressive symptoms (CSDSs) and ARHL in middle-aged and older adults using data from the China Health and Retirement Longitudinal Study. Among 3,418 participants free of baseline ARHL, baseline CSDS was associated with an increased odds of incident ARHL (odds ratio [OR]: 1.51). Cognitive decline, BMI, and arthritis partially mediated the longitudinal CSDS–ARHL association and explained 24% of the variance in the total effect. Among 4,921 participants without baseline CSDS, baseline ARHL was associated with an increased odds of incident CSDS (OR: 1.37). The bidirectional associations remained significant after adjustments for baseline demographic factors, comorbidities, and other health-related covariates. Depression may contribute to the development of ARHL, and vice versa. Interventions in depression, cognitive decline, and arthritis may delay the onset of ARHL and break the vicious circle between them.

scholarly journals A simple new method for identifying performance characteristics associated with success in elite tennis

2018 ◽  
Vol 14 (1) ◽  
pp. 43-50 ◽  
Author(s):  
Anna Fitzpatrick ◽  
Joseph A Stone ◽  
Simon Choppin ◽  
John Kelley
Keyword(s):  
Performance Analysis ◽  
Statistical Methods ◽  
Rank Order ◽  
Large Datasets ◽  
Performance Characteristics ◽  
New Method ◽  
Considerable Time ◽  
Simple Method ◽  
Suitable Alternative ◽  
Using Data

Performance analysis and identifying performance characteristics associated with success are of great importance to players and coaches in any sport. However, while large amounts of data are available within elite tennis, very few players employ an analyst or attempt to exploit the data to enhance their performance; this is partly attributable to the considerable time and complex techniques required to interpret these large datasets. Using data from the 2016 and 2017 French Open tournaments, we tested the agreement between the results of a simple new method for identifying important performance characteristics (the Percentage of matches in which the Winner Outscored the Loser, PWOL) and the results of two standard statistical methods to establish the validity of the simple method. Spearman’s rank-order correlations between the results of the three methods demonstrated excellent agreement, with all methods identifying the same three performance characteristics ( points won of 0–4 rally length, baseline points won and first serve points won) as strongly associated with success. Consequently, we propose that the PWOL method is valid for identifying performance characteristics associated with success in tennis, and is therefore a suitable alternative to more complex statistical methods, as it is simpler to calculate, interpret and contextualise.

scholarly journals Loneliness effects on personality

2018 ◽  
Vol 43 (2) ◽  
pp. 136-146 ◽  
Author(s):  
Marcus Mund ◽  
Franz J. Neyer
Keyword(s):  
Social Relationships ◽  
Big Five ◽  
Regulatory Focus ◽  
Statistical Significance ◽  
Negative Events ◽  
Birth Cohorts ◽  
German Sample ◽  
Health Related ◽  
Using Data

Individuals feel lonely when they perceive a discrepancy between their desired and their actually experienced quantity and quality of social relationships. Prior research has demonstrated the importance of loneliness for various health-related aspects. In the present article, we extend the existing literature on loneliness by investigating its role for predicting personality traits and their development from late adolescence to early midlife. Using data from a representative German sample ( N = 12,402) sampling individuals from three different birth cohorts, we found loneliness to predict the levels of all Big Five traits except openness five years later. The effects of loneliness on the development of neuroticism and extraversion reached statistical significance but were only marginal in terms of effect size. Furthermore, we found that a self-regulatory focus geared to the prevention of negative events mediated the effects of loneliness on later levels of the Big Five.

Sign in / Sign up

Export Citation Format

Share Document