Semi-supervised Estimation Research Articles

Segmentation-based cardiomegaly detection based on semi-supervised estimation of cardiothoracic ratio

The successful integration of neural networks in a clinical setting is still uncommon despite major successes achieved by artificial intelligence in other domains. This is mainly due to the black box characteristic of most optimized models and the undetermined generalization ability of the trained architectures. The current work tackles both issues in the radiology domain by focusing on developing an effective and interpretable cardiomegaly detection architecture based on segmentation models. The architecture consists of two distinct neural networks performing the segmentation of both cardiac and thoracic areas of a radiograph. The respective segmentation outputs are subsequently used to estimate the cardiothoracic ratio, and the corresponding radiograph is classified as a case of cardiomegaly based on a given threshold. Due to the scarcity of pixel-level labeled chest radiographs, both segmentation models are optimized in a semi-supervised manner. This results in a significant reduction in the costs of manual annotation. The resulting segmentation outputs significantly improve the interpretability of the architecture’s final classification results. The generalization ability of the architecture is assessed in a cross-domain setting. The assessment shows the effectiveness of the semi-supervised optimization of the segmentation models and the robustness of the ensuing classification architecture.

Semi-supervised estimation for the varying coefficient regression model

<abstract><p>In many cases, the 'labeled' outcome is difficult to observe and may require a complicated or expensive procedure, and the predictor information is easy to be obtained. We propose a semi-supervised estimator for the one-dimensional varying coefficient regression model which improves the conventional supervised estimator by using the unlabeled data efficiently. The semi-supervised estimator is proposed by introducing the intercept model and its asymptotic properties are proven. The Monte Carlo simulation studies and a real data example are conducted to examine the finite sample performance of the proposed procedure.</p></abstract>

Semi-supervised estimation of capacity degradation for lithium ion batteries with electrochemical impedance spectroscopy

Machine learning-based methods have emerged as a promising solution to accurate battery capacity estimation for battery management systems. However, they are generally developed in a supervised manner which requires a considerable number of input features and corresponding capacities, leading to prohibitive costs and efforts for data collection. In response to this issue, this study proposes a convolutional neural network (CNN) based method to perform end-to-end capacity estimation by taking only raw impedance spectra as input. More importantly, an input reconstruction module is devised to effectively exploit impedance spectra without corresponding capacities in the training process, thereby significantly alleviating the cost of collecting training data. Two large battery degradation datasets encompassing over 4700 impedance spectra are developed to validate the proposed method. The results show that accurate capacity estimation can be achieved when substantial training samples with measured capacities are given. However, the estimation performance of supervised machine learning algorithms sharply deteriorates when fewer samples with measured capacities are available. In this case, the proposed method outperforms supervised benchmarks and can reduce the root mean square error by up to 50.66%. A further validation under different current rates and states of charge confirms the effectiveness of the proposed method. Our method provides a flexible approach to take advantage of unlabelled samples for developing data-driven models and is promising to be generalised to other battery management tasks.

Robust and efficient semi-supervised estimation of average treatment effects with application to electronic health records data.

We consider the problem of estimating the average treatment effect (ATE) in a semi-supervised learning setting, where a very small proportion of the entire set of observations are labeled with the true outcome but features predictive of the outcome are available among all observations. This problem arises, for example, when estimating treatment effects in electronic health records (EHR) data because gold-standard outcomes are often not directly observable from the records but are observed for a limited number of patients through small-scale manual chart review. We develop an imputation-based approach for estimating the ATE that is robust to misspecification of the imputation model. This effectively allows information from the predictive features to be safely leveraged to improve efficiency in estimating the ATE. The estimator is additionally doubly-robust in that it is consistent under correct specification of either an initial propensity score model or a baseline outcome model. It is also locally semiparametric efficient under an ideal semi-supervised model where the distribution of the unlabeled data is known. Simulations exhibit the efficiency and robustness of the proposed method compared to existing approaches in finite samples. We illustrate the method by comparing rates of treatment response to two biologic agents for treatment inflammatory bowel disease using EHR data from Partners'Healthcare.

Semi-supervised estimation of covariance with application to phenome-wide association studies with electronic medical records data.

Electronic medical records data are valuable resources for discovery research. They contain detailed phenotypic information on individual patients, opening opportunities for simultaneously studying multiple phenotypes. A useful tool for such simultaneous assessment is the phenome-wide association study, which relates a genomic or biological marker of interest to a wide spectrum of disease phenotypes, typically defined by the diagnostic billing codes. One challenge arises when the biomarker of interest is expensive to measure on the entire electronic medical record cohort. Performing phenome-wide association study based on supervised estimation using only subjects who have marker measurements may yield limited power. In this paper, we focus on the setting where the marker is measured on a small fraction of the patients while a few surrogate markers such as historical measurements of the biomarker are available on a large number of patients. We propose an efficient semi-supervised estimation procedure to estimate the covariance between the biomarker and the billing code, leveraging the surrogate marker information. We employ surrogate marker values to impute the missing outcome via a two-step semi-non-parametric approach and demonstrate that our proposed estimator is always more efficient than the supervised counterpart without requiring the imputation model to be correct. We illustrate the proposed procedure by assessing the association between the C-reactive protein and some inflammatory diseases with an electronic medical record study of inflammatory bowel disease performed with the Partners HealthCare electronic medical record database where C-reactive protein was only measured for a small fraction of the patients due to budget constraints.

Inverse scattering in a Bayesian framework: application to microwave imaging for breast cancer detection

In this paper, we deal with a nonlinear inverse scattering problem where the goal is to detect breast cancer from measurements of the scattered field that results from the interaction between the breast and a known interrogating wave in the microwave frequency range. Modeling of the wave-object (breast) interaction is tackled through a domain integral representation of the electric field in a 2D-TM configuration. The inverse problem is solved in a Bayesian framework where prior information, which consists in the fact that the object is supposed to be composed of compact homogeneous regions made of a restricted number of different materials, is introduced via a Gauss−Markov−Potts model. As an analytic expression for the joint maximum a posteriori (MAP) estimators yields an intractable solution, an approximation of the latter is proposed. This is done by means of a variational Bayesian approximation (VBA) technique that is adapted to complex-valued contrast and applied to compute the posterior estimators, and reconstruct maps of both permittivity and conductivity of the sought object. This leads to a joint semi-supervised estimation approach, which allows us to estimate the induced currents, the contrast and all of the parameters introduced in the prior model. The method is tested on two sets of synthetic data generated in different configurations and its performances are compared to that given by a contrast source inversion technique.