Interactive Q-learning for Quantiles.
Ontology highlight
ABSTRACT:
SUBMITTER: Linn KA
PROVIDER: S-EPMC5586239 | biostudies-literature | 2017
REPOSITORIES: biostudies-literature
Publications
Interactive Q-learning for Quantiles.
Linn Kristin A KA Laber Eric B EB Stefanski Leonard A LA
Journal of the American Statistical Association 20170331 518
A dynamic treatment regime is a sequence of decision rules, each of which recommends treatment based on features of patient medical history such as past treatments and outcomes. Existing methods for estimating optimal dynamic treatment regimes from data optimize the mean of a response variable. However, the mean may not always be the most appropriate summary of performance. We derive estimators of decision rules for optimizing probabilities and quantiles computed with respect to the response dis ...[more]
Similar Datasets
Project description:Quantiles and percentiles represent useful statistical tools for describing the distribution of results and deriving reference intervals and performance specification in laboratory medicine. They are commonly intended as the sample estimate of a population parameter and therefore they need to be presented with a confidence interval (CI). In this work we discuss three methods to estimate CI on quantiles and percentiles using parametric, nonparametric and resampling (bootstrap) approaches. The result of our numerical simulations is that parametric methods are always more accurate regardless of sample size when the procedure is appropriate for the distribution of results for both extreme (2.5th and 97.5th) and central (25th, 50th and 75th) percentiles and corresponding quantiles. We also show that both nonparametric and bootstrap methods suit well the CI of central percentiles that are used to derive performance specifications through quality indicators of laboratory processes whose underlying distribution is unknown.
| S-EPMC6294150 | biostudies-literature
Project description:Evidence-based rules for optimal treatment allocation are key components in the quest for efficient, effective health care delivery. Q-learning, an approximate dynamic programming algorithm, is a popular method for estimating optimal sequential decision rules from data. Q-learning requires the modeling of nonsmooth, nonmonotone transformations of the data, complicating the search for adequately expressive, yet parsimonious, statistical models. The default Q-learning working model is multiple linear regression, which is not only provably misspecified under most data-generating models, but also results in nonregular regression estimators, complicating inference. We propose an alternative strategy for estimating optimal sequential decision rules for which the requisite statistical modeling does not depend on nonsmooth, nonmonotone transformed data, does not result in nonregular regression estimators, is consistent under a broader array of data-generation models than Q-learning, results in estimated sequential decision rules that have better sampling properties, and is amenable to established statistical approaches for exploratory data analysis, model building, and validation. We derive the new method, IQ-learning, via an interchange in the order of certain steps in Q-learning. In simulated experiments IQ-learning improves on Q-learning in terms of integrated mean squared error and power. The method is illustrated using data from a study of major depressive disorder.
| S-EPMC4274394 | biostudies-literature
Project description:BackgroundBlended learning has the potential to stimulate effective learning and to facilitate high quality education. For postgraduate health professionals, blended learning is relatively new. For this group we developed, implemented and evaluated two blended learning modules in a master program on quality and safety in patient care. Aiming for a better preparation compared to traditional textbook homework, the course provided not only web lectures and reading, but also interactive assignments and collaborative learning. Additional goal was saving time for the teachers resulting in a potential cost savings.MethodThe experiences of 21 postgraduate health professionals were evaluated with two voluntary and anonymous questionnaires beginning of 2017 with a special focus on the added value of online interaction, underexposed in previous research.ResultsThis evaluation shows that online modules are regarded as being an effective preparation for face-to-face meetings for postgraduate health professionals. Added value of social interactive online preparation was perceived from collaborating and interacting with each other. Both the interaction between the students, and the e-moderator and teachers were well received.ConclusionsBased on the results of this study, we suggest that blended learning may indeed increase the level of education and stimulate effective learning for postgraduate health care professionals. The professionals experienced added value of social interactive online preparation from collaborating and interacting with each other. We consider better aligning of the online and face-to-face components as one of the highest priorities.
| S-EPMC6664728 | biostudies-literature
Project description:ObjectiveMedical word sense disambiguation (WSD) is challenging and often requires significant training with data labeled by domain experts. This work aims to develop an interactive learning algorithm that makes efficient use of expert's domain knowledge in building high-quality medical WSD models with minimal human effort.MethodsWe developed an interactive learning algorithm with expert labeling instances and features. An expert can provide supervision in 3 ways: labeling instances, specifying indicative words of a sense, and highlighting supporting evidence in a labeled instance. The algorithm learns from these labels and iteratively selects the most informative instances to ask for future labels. Our evaluation used 3 WSD corpora: 198 ambiguous terms from Medical Subject Headings (MSH) as MEDLINE indexing terms, 74 ambiguous abbreviations in clinical notes from the University of Minnesota (UMN), and 24 ambiguous abbreviations in clinical notes from Vanderbilt University Hospital (VUH). For each ambiguous term and each learning algorithm, a learning curve that plots the accuracy on the test set against the number of labeled instances was generated. The area under the learning curve was used as the primary evaluation metric.ResultsOur interactive learning algorithm significantly outperformed active learning, the previous fastest learning algorithm for medical WSD. Compared to active learning, it achieved 90% accuracy for the MSH corpus with 42% less labeling effort, 35% less labeling effort for the UMN corpus, and 16% less labeling effort for the VUH corpus.ConclusionsHigh-quality WSD models can be efficiently trained with minimal supervision by inviting experts to label informative instances and provide domain knowledge through labeling/highlighting contextual features.
| S-EPMC6658868 | biostudies-literature
Project description:IntroductionThis resource is a web-based and self-contained self-learning module (SLM) on binge-eating disorder (BED) that runs utilizing Articulate software. Certain skills and knowledge are required to both interview and manage patients diagnosed with the disorder. Because BED may be encountered by medical students, this module can help inform the learning process.MethodsSoftware was utilized to develop an SLM for third-year medical students utilizing video clips demonstrating the evolution of a clinical BED case from presentation to follow-up over 10-12 weeks. The focus is on differential diagnosis and management, with intermittent self-assessment quizzes and other evidence-based literature for further study. The module follows the case of Ms. Melanie May, who has been referred for help to lose weight. Further assessment reveals that she has BED. By having her needs therapeutically attended to, she receives help through focused cognitive behavioral therapy, the use of food log monitoring, and medications.ResultsThis SLM on BED has been implemented at the University of Toronto as an optional component of the psychiatry clerkship over the 2015/2016 academic year. Written and verbal feedback from faculty (n = 5) and students (n = 5) has been overwhelmingly positive.DiscussionThis SLM is intended as a clinical application exercise for third-year medical students who may not encounter patients with BED or who want to increase their clinical understanding of how to help patients who present with obesity and binge-eating complaints.
| S-EPMC6342049 | biostudies-literature
Project description:IntroductionPediatrics is an exciting field because it requires having a background in physiology that evolves as the patient ages. As a subspecialty, pediatric endocrinology encompasses a wide range of disease processes both acute and chronic. This module was created to provide a review of endocrine physiology, promote understanding of the biopsychosocial model of children diagnosed with an endocrine disorder, and utilize simulated case studies to become familiar with patient management.MethodsThe material and case studies presented are based on firsthand experiences in a pediatric clinic and diabetes camp, as well as an extensive literature review. The information in this resource teaches interview questions and problem-solving techniques and ensures that the learner understands the basic equipment used by the patients. Implementation of this module will advance students' and residents' efficacy in caring for this pediatric population. The approximate time to complete this module is 3 to 5 hours. To evaluate the effectiveness of this module, pre- and postmodule surveys were administered to medical students. Factors analyzed included overall user satisfaction, utility of the module, comfort in approaching pediatric endocrine patients, and suggestions for improvement.ResultsA cohort of medical students (N = 26) completed both surveys and the module, with an equal distribution of first- and second-year students to third- and fourth-year students. The surveys showed a general trend of improvement in self-reported comfort with both basic science and clinical skills components.DiscussionThis module would be most beneficial to a medical student who may rotate with a pediatric endocrinologist during the clinical years. The target audience can be broadened to include not only medical students but also physician assistant students, nursing/nurse practitioner students, pediatric residents, and pediatric endocrinology fellows.
| S-EPMC6464428 | biostudies-literature
Project description:Automated profiling of cell morphology is a powerful tool for inferring cell function. However, this technique retains a high barrier to entry. In particular, configuring image processing parameters for optimal cell profiling is susceptible to cognitive biases and dependent on user experience. Here, we use interactive machine learning to identify the optimum cell profiling configuration that maximises quality of the cell profiling outcome. The process is guided by the user, from whom a rating of the quality of a cell profiling configuration is obtained. We use Bayesian optimisation, an established machine learning algorithm, to learn from this information and automatically recommend the next configuration to examine with the aim of maximising the quality of the processing or analysis. Compared to existing interactive machine learning tools that require domain expertise for per-class or per-pixel annotations, we rely on users' explicit assessment of output quality of the cell profiling task at hand. We validated our interactive approach against the standard human trial-and-error scheme to optimise an object segmentation task using the standard software CellProfiler. Our toolkit enabled rapid optimisation of an object segmentation pipeline, increasing the quality of object segmentation over a pipeline optimised through trial-and-error. Users also attested to the ease of use and reduced cognitive load enabled by our machine learning strategy over the standard approach. We envision that our interactive machine learning approach can enhance the quality and efficiency of pipeline optimisation to democratise image-based cell profiling.
| S-EPMC7485821 | biostudies-literature
Project description:We demonstrate that questions of convergence and divergence regarding shapes of distributions can be carried out in a location- and scale-free environment. This environment is the class of probability density quantiles (pdQs), obtained by normalizing the composition of the density with the associated quantile function. It has earlier been shown that the pdQ is representative of a location-scale family and carries essential information regarding shape and tail behavior of the family. The class of pdQs are densities of continuous distributions with common domain, the unit interval, facilitating metric and semi-metric comparisons. The Kullback-Leibler divergences from uniformity of these pdQs are mapped to illustrate their relative positions with respect to uniformity. To gain more insight into the information that is conserved under the pdQ mapping, we repeatedly apply the pdQ mapping and find that further applications of it are quite generally entropy increasing so convergence to the uniform distribution is investigated. New fixed point theorems are established with elementary probabilistic arguments and illustrated by examples.
| S-EPMC7512836 | biostudies-literature
Project description:New computer vision solutions combined with artificial intelligence algorithms can help recognize patterns in biological images, reducing subjectivity and optimizing the analysis process. The aim of this study was to propose an approach based on interactive and traditional machine learning methods to classify soybean seeds and seedlings according to their appearance and physiological potential. In addition, we correlated the appearance of seeds to their physiological performance. Images of soybean seeds and seedlings were used to develop models using low-cost approaches and free-access software. The models developed showed high performance, with overall accuracy reaching 0.94 for seeds and seedling classification. The high precision of the models that were developed based on interactive and traditional machine learning demonstrated that the method can easily be used to classify soybean seeds according to their appearance, as well as to classify soybean seedling vigor quickly and non-subjectively. The appearance of soybean seeds is strongly correlated with their physiological performance.
| S-EPMC7347887 | biostudies-literature
Project description:Traditional teaching methods in radiology education have not kept pace with advances in technology that foster successful transition into independent practice. This deficit has been exacerbated by the COVID-19 pandemic, as the need for social distancing and the introduction of hybrid staffing models have decreased the critical educational interactions at the reading room workstations between staff and trainees. By leveraging interactive, case-based learning, educators have the opportunity to bridge the substantial gap between basic pattern recognition and successfully making a diagnosis in independent practice. For the educator, this signals a shift away from perfect case selection and presenter authority, and toward the role of a guide facilitating an active case-based learning experience. This form of learning is best accompanied by guided interpretation and iterative feedback with the goal of developing similar levels of mastery and autonomy among graduating trainees. In this article, we present the tools and methods for incorporating interactive cases into existing and novel teaching materials to meet the unique challenges educators are facing today.
| S-EPMC8175917 | biostudies-literature