Information Theoretic Learning-Enhanced Dual-Generative Adversarial Networks With Causal Representation for Robust OOD Generalization.

Recently, machine/deep learning techniques are achieving remarkable success in a variety of intelligent control and management systems, promising to change the future of artificial intelligence (AI) scenarios. However, they still suffer from some intractable difficulty or limitations for model training, such as the out-of-distribution (OOD) issue, in modern smart manufacturing or intelligent transportation systems (ITSs). In this study, we newly design and introduce a deep generative model framework, which seamlessly incorporates the information theoretic learning (ITL) and causal representation learning (CRL) in a dual-generative adversarial network (Dual-GAN) architecture, aiming to enhance the robust OOD generalization in modern machine learning (ML) paradigms. In particular, an ITL-and CRL-enhanced Dual-GAN (ITCRL-DGAN) model is presented, which includes an autoencoder with CRL (AE-CRL) structure to aid the dual-adversarial training with causality-inspired feature representations and a Dual-GAN structure to improve the data augmentation in both feature and data levels. Following a newly designed feature separation strategy, a causal graph is built and improved based on the information theory, which can enhance the causally related factors among the separated core features and further enrich the feature representation with the counterfactual features via interventions based on the refined causal relationships. The ITL is incorporated to improve the extraction of low-dimensional feature representations and learn the optimized causal representations based on the idea of "information flow." A dual-adversarial training mechanism is then developed, which not only enables the generator to expand the boundary of feature distribution in accordance with the optimized feature representation from AE-CRL, but also allows the discriminator to further verify and improve the quality of the augmented data for OOD generalization. Experiment and evaluation results based on an open-source dataset demonstrate the outstanding learning efficiency and classification performance of our proposed model for robust OOD generalization in modern smart applications compared with three baseline methods.

Pneumothorax with Eosinophilia is an Important Diagnostic Clue for Distinguishing Paragonimiasis from Chronic Eosinophilic Pneumonia: A Case Report.

The Paragonimus westermani infection is a parasitic foodborne infection that induces systemic symptoms with eosinophilia in humans. Here, we described pneumothorax in addition to pulmonary opacities with eosinophilia in a man with a positive P. westermani serology. He was misdiagnosed with chronic eosinophilic pneumonia (CEP) during the initial phase. Paragonimiasis can share similar clinical findings with CEP in cases where the worm is confined to the lungs. The findings of the current study suggest that paragonimiasis and CEP can be distinguished from each other by the presence of various symptoms. Notably, eosinophilia with pneumothorax should be an important diagnostic factor for paragonimiasis.

Force-Induced Shuttling of Rotaxanes Controls Fluorescence Resonance Energy Transfer in Polymer Hydrogels.

The molecular shuttling function of rotaxanes can be exploited to design mechanoresponsive reporter molecules. Here, we report a new approach to such rotaxane-based mechanophores, in which the fluorescence resonance energy transfer (FRET) between a donor-acceptor pair is mechanically controlled. A cyclic molecule containing a green-light-emitting FRET donor connected to a red-light-emitting FRET acceptor was threaded onto an axle equipped with a quencher at its center and two stoppers in the peripheral positions. In the force-free state, the green emitter is located near the quencher so that charge transfer interactions or photo-induced electron transfer between the two moieties suppress green emission and prevent the FRET from the green to the red emitter. The mechanophore was covalently incorporated into a linear polyurethane-urea (PUU), and stretchable hydrogels were prepared by swelling this polymer with water. Upon deformation of the PUU hydrogels and under an excitation light that selectively excites the donor, the intensity of the red fluorescence increases, as a result of a force-induced separation of the green emitter from the quencher, which enables the FRET. The switching contrast is much more pronounced in the gels than in dry films, which is due to increased molecular mobility and hydrophobic effects in the hydrogel, which both promote the formation of inclusion complexes between the ring containing the green emitter and the quencher.

Nonlinear Causal Discovery for High-Dimensional Deterministic Data.

Nonlinear causal discovery with high-dimensional data where each variable is multidimensional plays a significant role in many scientific disciplines, such as social network analysis. Previous work majorly focuses on exploiting asymmetry in the causal and anticausal directions between two high-dimensional variables (a cause-effect pair). Although there exist some works that concentrate on the causal order identification between multiple variables, i.e., more than two high-dimensional variables, they do not validate the consistency of methods through theoretical analysis on multiple-variable data. In particular, based on the asymmetry for the cause-effect pair, if model assumptions for any pair of the data are violated, the asymmetry condition will not hold, resulting in the deduction of incorrect order identification. Thus, in this article, we propose a causal functional model, namely high-dimensional deterministic model (HDDM), to identify the causal orderings among multiple high-dimensional variables. We derive two candidates' selection rules to alleviate the inconvenient effects resulted from the violated-assumption pairs. The corresponding theoretical justification is provided as well. With these theoretical results, we develop a method to infer causal orderings for nonlinear multiple-variable data. Simulations on synthetic data and real-world data are conducted to verify the efficacy of our proposed method. Since we focus on deterministic relations in our method, we also verify the robustness of the noises in simulations.

In-hospital blood collection increases the rate of indeterminate results in interferon-gamma release assays.

BACKGROUND: The interferon (IFN)-γ release assay (IGRA) has recently been established as a method to evaluate the infection status of tuberculosis instead of the tuberculin skin test. However, indeterminate results can create challenges to interpretation. The IGRA has been available in Japan since 2005, including the recently launched QuantiFERON-TB Gold Plus (QFT-plus) assay. OBJECTIVES: The aim of this study was to investigate the clinical features and predictors of indeterminate results by the QFT-plus test in routine practice. METHODS: This was a cross-sectional study of 1258 patients. Multivariate logistic regression models were employed to investigate the clinical factors related to indeterminate results by the QFT-plus. RESULTS: Overall, 91.8% of results were found to be conclusive and 8.2% were indeterminate. The QFT-plus indeterminate results were predominantly due to a low level of IFN-γ production by mitogens. Multivariate analysis indicated that an indeterminate result was significantly associated with age, sex, corticosteroid use, autoimmune disease, and inpatient setting. CONCLUSION: Certain types of individuals are at higher risk of an indeterminate IGRA result. The QFT-plus test for hospitalized patients should be avoided as much as possible, and it is better to perform the test for those patients in outpatient settings.

Chemical-Mediated Microbial Interactions Can Reduce the Effectiveness of Time-Series-Based Inference of Ecological Interaction Networks.

Network-based assessments are important for disentangling complex microbial and microbial-host interactions and can provide the basis for microbial engineering. There is a growing recognition that chemical-mediated interactions are important for the coexistence of microbial species. However, so far, the methods used to infer microbial interactions have been validated with models assuming direct species-species interactions, such as generalized Lotka-Volterra models. Therefore, it is unclear how effective existing approaches are in detecting chemical-mediated interactions. In this paper, we used time series of simulated microbial dynamics to benchmark five major/state-of-the-art methods. We found that only two methods (CCM and LIMITS) were capable of detecting interactions. While LIMITS performed better than CCM, it was less robust to the presence of chemical-mediated interactions, and the presence of trophic competition was essential for the interactions to be detectable. We show that the existence of chemical-mediated interactions among microbial species poses a new challenge to overcome for the development of a network-based understanding of microbiomes and their interactions with hosts and the environment.

Impact of the interferon-γ release assay and glomerular filtration rate on the estimation of active tuberculosis risk before bronchoscopic examinations: a retrospective pilot study.

BACKGROUND: Bronchoscopic examinations are vital to diagnose pulmonary diseases. However, as coughing is triggered during and after the procedure, it is imperative to take measures against nosocomial infections, especially for airborne infections like tuberculosis (TB). The interferon-γ release assay (IGRA) has recently been established as a method to evaluate the infection status of TB. We aimed to ascertain the efficacy of IGRA and clinical findings in estimating the prevalence of active TB before bronchoscopy. METHODS: We obtained IGRA results from 136 inpatients using a QuantiFERON-TB Gold In-Tube test. Bronchoscopy samples were cultured in Mycobacteria Growth indicator tubes and 2% Ogawa solid medium. We evaluated the adjusted effects of multiple clinical variables on active TB status using a logistic regression model. In addition, multiple variables were converted into a decision tree to predict active TB. RESULTS: Five (3.7%) patients were diagnosed with culture-positive TB, two of whom were simultaneously diagnosed with non-small-cell lung carcinoma or small-cell lung carcinoma. The multivariate analysis suggested the probability of predicting active TB using the IGRA [odds ratio (OR), 72.7; 95% confidence interval (CI), 3.169-1668; P=0.007] and decreased estimated glomerular filtration rate (eGFR) (OR, 0.937; 95% CI, 0.882-0.996; P=0.038) in patients undergoing bronchoscopy. A decision tree validated the use of these two variables to predict active TB. CONCLUSIONS: IGRA test results are useful for predicting active TB before bronchoscopy. This strategy could identify patients who require antibiotic therapy to prevent TB or who are in the active phase of TB.

Analysis of cause-effect inference by comparing regression errors.

We address the problem of inferring the causal direction between two variables by comparing the least-squares errors of the predictions in both possible directions. Under the assumption of an independence between the function relating cause and effect, the conditional noise distribution, and the distribution of the cause, we show that the errors are smaller in causal direction if both variables are equally scaled and the causal relation is close to deterministic. Based on this, we provide an easily applicable algorithm that only requires a regression in both possible causal directions and a comparison of the errors. The performance of the algorithm is compared with various related causal inference methods in different artificial and real-world data sets.

Non-Gaussian Methods for Causal Structure Learning.

Causal structure learning is one of the most exciting new topics in the fields of machine learning and statistics. In many empirical sciences including prevention science, the causal mechanisms underlying various phenomena need to be studied. Nevertheless, in many cases, classical methods for causal structure learning are not capable of estimating the causal structure of variables. This is because it explicitly or implicitly assumes Gaussianity of data and typically utilizes only the covariance structure. In many applications, however, non-Gaussian data are often obtained, which means that more information may be contained in the data distribution than the covariance matrix is capable of containing. Thus, many new methods have recently been proposed for using the non-Gaussian structure of data and inferring the causal structure of variables. This paper introduces prevention scientists to such causal structure learning methods, particularly those based on the linear, non-Gaussian, acyclic model known as LiNGAM. These non-Gaussian data analysis tools can fully estimate the underlying causal structures of variables under assumptions even in the presence of unobserved common causes. This feature is in contrast to other approaches. A simulated example is also provided.

Meiotic chromosomal recombination defect in sake yeasts.

Sake yeast strains are classified into Saccharomyces cerevisiae and have a heterothallic life cycle. This feature allows cross hybridization between two haploids to breed new strains with superior characteristics. However, cross hybridization of sake yeast is very difficult because only a few spores develop in a sporulation medium, and most of these spores do not germinate. We hypothesized that these features are attributable to chromosome recombination defect in meiosis, which leads to chromosome loss. To test this hypothesis, we examined meiotic recombination of sake yeast Kyokai no. 7 (K7) using the following three methods: (i) analysis of the segregation patterns of two heterozygous sites in the same chromosome in 100 haploid K7 strains; (ii) sequencing of the whole genomes of four haploid K7 strains and comparison of the bases derived from the heterozygosities; and (iii) construction of double heterozygous disruptants of CAN1 and URA3 on the chromosome V of K7 and the examination of the genotypes of haploids after sporulation. We could not detect any recombinant segregants in any of the experiments, which indicated defect in meiotic recombination in K7. Analyses after sporulation of the same double heterozygous disruptants of K6, K9, and K10 also indicated meiotic recombination defect in these strains. Although rapamycin treatment increased the sporulation efficiency of K7, it did not increase the meiotic recombination of the double heterozygous K7. Moreover, the spo13 disruptant of the K7 derivative produced two spore asci without meiotic recombination. These results suggest that sake yeasts have defects in meiotic recombination machinery.

[PROBLEMS ASSOCIATED WITH THE DIAGNOSIS OF PROLONGED/CHRONIC COUGH IN CLINICAL SETTINGS].

BACKGROUND: Many patients visit primary care clinics or local hospitals with a complaint of prolonged/chronic cough. Among the different types of chronic cough, cough variant asthma (CVA) and postinfectious cough may be the most common types, and their differential diagnosis is difficult. Some physicians tend to prescribe inhaled corticosteroids before establishing a definitive diagnosis. METHODS: We retrospectively investigated useful findings for diagnosis in 77 patients with a complaint of prolonged/ chronic cough to detect meaningful findings for differential diagnosis and to identify problems associated with diagnosis in clinical practice. RESULTS: CVA was diagnosed in 39 patients, and postinfectious cough was diagnosed in 19. Compared with postinfectious cough, CVA was associated with significant characteristics such as "diurnal variation of symptoms," "response to inhalation of short acting ß2 agonist (SABA)," and "recurrent episodes of symptoms." CVA was associated with high FeNO levels as well, and high FeNO levels were specific to CVA. However, these useful characteristics were not significant in the patients who had been prescribed ICS before visiting our hospital. CONCLUSIONS: Medical examination and determination of FeNO levels are useful for the differential diagnosis of prolonged/chronic cough, before treatment with ICS.

Hepatocellular adenoma in a woman who was undergoing testosterone treatment for gender identity disorder.

A 32-year-old Japanese woman was admitted to our hospital for the diagnosis and treatment of multiple liver tumors. She had been receiving 125 mg testosterone enanthate every 2 weeks following female-to-male gender identity disorder (GID) diagnosis at 20 years of age. Ultrasonography, computed tomography, and magnetic resonance imaging showed 11 hepatic nodular tumors with a maximum diameter of 28 mm. Liver tumors with hepatocellular adenoma (HCA) were diagnosed with needle biopsy. Segmentectomy of the left lateral lobe including two lesions, subsegmentectomy of S6 including two lesions, enucleation of each tumor in S5 and S7, and open surgical radiofrequency ablation for each tumor in S4 and S7 were performed. Immunohistochemical specimens showed that the tumor cells were diffusely and strongly positive for glutamine synthetase and that the nuclei were ectopically positive for ß-catenin. Thus, the tumors were diagnosed as ß-catenin-activated HCA (b-HCA). Transcatheter arterial chemoembolization plus subsequent radiofrequency ablation was performed for the 3 residual lesions in S4 and S8. Although testosterone enanthate was being continued for GID, no recurrence was observed until at least 22 months after the intensive treatments. HCA development in such patients receiving testosterone should be closely monitored using image inspection.

C-Arm Cone Beam Computed Tomography Guidance for Radiofrequency Ablation in Hepatocellular Carcinoma.

OBJECTIVE: To assess the usefulness of C-arm cone beam computed tomography (CBCT) combined with ultrasound for the treatment of hepatocellular carcinoma (HCC) by radiofrequency ablation (RFA). METHODS: Patients underwent RFA following transcatheter arterial chemoembolization (TACE) or RFA alone under ultrasound or CBCT guidance combined with ultrasound-based techniques. They were divided into 2 groups based on the use (C group) and nonuse (NC group) of CBCT guidance. The technical success of RFA and local tumor progression after the first RFA session were evaluated by dynamic contrast-enhanced imaging methods. Between-group differences were assessed retrospectively. RESULTS: We enrolled 198 patients with 260 HCC nodules. The complete ablation rates were 63.0 and 89.4% in the NC and C groups, respectively. In log-rank testing, local tumor progression occurred significantly more often in the NC group when RFA was used without TACE, in males when des-gamma-carboxy prothrombin was ≥29 mAU/mL, and when the diameter of a nodule was ≥18 mm. On Cox proportional-hazards regression analysis, the NC group, RFA alone without TACE, and male gender were significant independent variables. CONCLUSION: TACE followed by RFA under CBCT and ultrasound guidance improves the reliability of ablation of target HCC nodules, reduces the need for additional treatment sessions, and prevents local tumor progression.

ParceLiNGAM: a causal ordering method robust against latent confounders.

We consider learning a causal ordering of variables in a linear nongaussian acyclic model called LiNGAM. Several methods have been shown to consistently estimate a causal ordering assuming that all the model assumptions are correct. But the estimation results could be distorted if some assumptions are violated. In this letter, we propose a new algorithm for learning causal orders that is robust against one typical violation of the model assumptions: latent confounders. The key idea is to detect latent confounders by testing independence between estimated external influences and find subsets (parcels) that include variables unaffected by latent confounders. We demonstrate the effectiveness of our method using artificial data and simulated brain imaging data.

Bayesian Estimation of Causal Direction in Acyclic Structural Equation Models with Individual-specific Confounder Variables and Non-Gaussian Distributions.

Several existing methods have been shown to consistently estimate causal direction assuming linear or some form of nonlinear relationship and no latent confounders. However, the estimation results could be distorted if either assumption is violated. We develop an approach to determining the possible causal direction between two observed variables when latent confounding variables are present. We first propose a new linear non-Gaussian acyclic structural equation model with individual-specific effects that are sometimes the source of confounding. Thus, modeling individual-specific effects as latent variables allows latent confounding to be considered. We then propose an empirical Bayesian approach for estimating possible causal direction using the new model. We demonstrate the effectiveness of our method using artificial and real-world data.

Estimating exogenous variables in data with more variables than observations.

Many statistical methods have been proposed to estimate causal models in classical situations with fewer variables than observations. However, modern datasets including gene expression data increase the needs of high-dimensional causal modeling in challenging situations with orders of magnitude more variables than observations. In this paper, we propose a method to find exogenous variables in a linear non-Gaussian causal model, which requires much smaller sample sizes than conventional methods and works even under orders of magnitude more variables than observations. Exogenous variables work as triggers that activate causal chains in the model, and their identification leads to more efficient experimental designs and better understanding of the causal mechanism. We present experiments with artificial data and real-world gene expression data to evaluate the method.

Role of antithrombotic therapy and nonsteroidal anti-inflammatory drug use in bleeding gastroduodenal ulcers.

OBJECTIVE: The aim of this study was to analyze the effects of anti-thrombotic therapy and nonsteroidal anti-inflammatory drugs (NSAIDs) on gastroduodenal bleeding. PATIENTS AND METHODS: The study subjects were 544 patients (421 males and 123 females, mean age, 64.2 years) who were treated endoscopically for bleeding gastroduodenal ulcers from January 1995 to August 2008. Of the 544 patients, 276 (50.7%) had a history of treatment for > or =1 month with an antithrombotic agent or NSAIDs, including low-dose aspirin (n=94), other NSAIDs (n=91), warfarin (n=43), or any combination of the three (combination treatment group; n=48). On the other hand, 268 patients had not previously received any of these drugs (control group). Clinical features and endoscopic therapeutic results were assessed and compared. RESULTS: Helicobacter pylori infection was detected in 187 of the 241 (77.6%) patients examined. Of the 544 patients, 199 (36.6%) attended the Department of Cardiology or Cardiovascular Surgery, and 170 (31.3%) patients were already being treated with antiulcer medication, including proton pump inhibitors (n=18 [3.3%]). Forty (7.4%) patients suffered from rebleeding after intervention; the incidence of ulcer rebleeding was not significantly different between patients being treated with any such drugs (4.7% [13/276]) and those that had not previously received any antithrombotic agent or NSAIDs (10.1% [27/268]). CONCLUSION: Antithrombotic therapy and NSAIDs use contributed to bleeding in 50.7% of patients with gastroduodenal ulcers. These drugs are a major cause of ulcer bleeding, but are not necessarily considered a risk factor for rebleeding after endoscopic hemostasis.