Search

Vision-and-language tasks have increasingly drawn more attention as a means to evaluate human-like reasoning in machine learning models. A popular task in the field is visual question answering (VQA), which aims to answer questions about images. However, VQA models have been shown to exploit language bias by learning the statistical correlations between questions and answers without looking into the image content: e.g., questions about the color of a banana are answered with yellow, even if the banana in the image is green. If societal bias (e.g., sexism, racism, ableism, etc.) is present in the training data, this problem may be causing VQA models to learn harmful stereotypes. For this reason, we investigate gender and racial bias in five VQA datasets. In our analysis, we find that the distribution of answers is highly different between questions about women and men, as well as the existence of detrimental gender-stereotypical samples. Likewise, we identify that specific race-related attributes are underrepresented, whereas potentially discriminatory samples appear in the analyzed datasets. Our findings suggest that there are dangers associated to using VQA datasets without considering and dealing with the potentially harmful stereotypes. We conclude the paper by proposing solutions to alleviate the problem before, during, and after the dataset collection process., Comment: ACM Conference on Fairness, Accountability, and Transparency (FAccT 2022)

BACKGROUND: Training for the fundus examination using traditional teaching is challenging, resulting in low generalist physicians’ confidence in performing the funduscopic exam. At the same time, there is growing evidence suggesting flipped classrooms’ value in teaching physical examination procedures. However, whether the flipped classroom is superior to the traditional, lecture-based teaching for the funduscopic exam and the cognitive processes supporting its effectiveness has not yet been determined. METHODS: We conducted a sequential explanatory mixed-method study to compare the flipped classroom approach’s effectiveness versus the traditional lecture-based classroom for teaching the funduscopic exam to the medical students at Chiba University in Japan. Medical students were randomly assigned to either a flipped classroom group or a traditional teaching group. We then quantitatively measured the diagnostic accuracy of funduscopic findings, the length of time to perform the fundus examination, and students’ confidence in performing funduscopic examinations, before and after attending the specific classrooms. Next, we conducted student focus groups to explore the students’ thinking processes in the flipped classroom and traditional teaching of fundus examination, respectively. The qualitative data were analyzed using the qualitative content analysis method.RESULTS: Diagnostic accuracy was significantly higher using the flipped classroom method (flipped: 36.6% to 63.4%, traditional: 28.3% to 34.6%, F (1,310) = 11.0, p = .001). The total examination time was significantly shorter using the flipped classroom teaching (flipped: 85.4s to 66.9s, traditional: 85.3s to 76.3s, F (1,310) = 14.7, p CONCLUSIONS: Teaching medical students funduscopic examination using the flipped classroom methodology leads to improved diagnostic accuracy, confidence, and motivation for funduscopic examinations, while reducing total examination time. The flipped classroom teaching method enabled higher levels of cognitive activity than the traditional, lecture-based classroom, as assessed using the revised Bloom’s taxonomy.

BackgroundTraining for the fundus examination using traditional teaching is challenging, resulting in low generalist physicians’ confidence in performing the funduscopic examination. There is growing evidence suggesting a flexible e-learning video approach’s value in teaching physical examination procedures. However, whether the flexible e-learning video approach is superior to the traditional, face-to-face (F2F) lecture-based teaching for the funduscopic exam and the cognitive processes supporting its effectiveness has not yet been determined.MethodsWe conducted a sequential explanatory mixed-method study to compare the flexible e-learning video approach’s effectiveness versus the F2F lecture-based approach for teaching the funduscopic exam to medical students at Chiba University in Japan. Medical students were randomly assigned to either a flexible e-learning video approach group or a F2F lecture approach group. We then quantitatively measured the diagnostic accuracy of funduscopic findings before and after attending the specific classrooms. Next, we conducted student focus groups to explore the students’ thinking processes in the flexible e-learning video approach vs. the F2F lecture-based teaching of fundus examination. The qualitative data were analyzed using the qualitative content analysis method.ResultsThe mean diagnostic accuracy scores in the post-test significantly increased from pre-test in the intervention group (36.6 to 63.4%,p n = 36). In the flexible e-learning video approach group, we identified ten categories corresponding to four levels of the revised Bloom’s taxonomy: remember, understand, apply, analyze. Five categories were identified in the traditional F2F lecture approach group corresponding to three revised Bloom’s taxonomy levels: understand, apply, analyze. Interrater reliability was substantial (Cohen’s kappa = 0.81).ConclusionsTeaching medical students funduscopic examination using the flexible e-learning video approach leads to improved diagnostic accuracy of funduscopic examinations. The flexible e-learning video teaching method enabled higher cognitive activity levels than the traditional, lecture-based classroom, as assessed using the revised Bloom’s taxonomy.Trial registrationThis study was registered with the University Hospital Medical Information Network Clinical Trials Registry on 08/02/2020 (Unique trial number:UMIN 000039434).

We investigate new methods for reliable provisioning of dynamic content requests in optical metro networks. Our methods leverage content replication across multiple edge datacenters and multipath routing.

We experimentally demonstrate cascaded performance degradation caused by increased inter-core crosstalk due to optical power fluctuation of adjacent core in multi-core fiber networks, and we show that a burst-mode EDFA is a candidate solution.

We discuss how different degrees of slice isolation influence resource allocation in protected optical metro-aggregation networks. The case of slice reliability with dedicated protection at lightpath is modelled and numerically evaluated.

Network connectivity, i.e., the reachability of any network node from all other nodes, is often considered as the default network survivability metric against failures. However, in the case of a large-scale disaster disconnecting multiple network components, network connectivity may not be achievable. On the other hand, with the shifting service paradigm towards the cloud in today’s networks, most services can still be provided as long as at least a content replica is available in all disconnected network partitions. As a result, the concept of content connectivity has been introduced as a new network survivability metric under a large-scale disaster. Content connectivity is defined as the reachability of content from every node in a network under a specific failure scenario. In this work, we investigate how to ensure content connectivity in optical metro networks. We derive necessary and sufficient conditions and develop what we believe to be a novel mathematical formulation to map a virtual network over a physical network such that content connectivity for the virtual network is ensured against multiple link failures in the physical network. In our numerical results, obtained under various network settings, we compare the performance of mapping with content connectivity and network connectivity and show that mapping with content connectivity can guarantee higher survivability, lower network bandwidth utilization, and significant improvement of service availability.

Type B insulin resistance (TBIR) is an extremely rare disease characterized by marked hyperglycemia and insulin resistance and often coexists with autoimmune diseases. The characteristics, symptoms, blood glucose patterns, comorbidities, and treatments of TBIR all vary and are not defined. In this study, we described a case of TBIR that developed 6 months after DPP-4 inhibitor administration and immediately after the patient caught a cold. Treatment using prednisolone and insulin-like growth factor-1 was effective. We also conducted an observational survey-based case series study in a Japanese cohort comprising 21 cases. The average age of onset of TBIR was 62.3±14.8 (17–84) years, and 61.9% of subjects were male. The majority of patients (90.4%) were 50 years old and over. During the study period, there was a high percentage (85.7%) of episodes of hypoglycemia, which was the trigger for diagnosis in more than 50% of cases. Glycemic patterns included 7 cases of hyperglycemia (33.3%), 10 cases of hypoglycemia (47.6%), and 4 cases of both hyperglycemia and hypoglycemia (19.1%). In the hypoglycemic group, 90.0% of patients were male. Furthermore, 71.4% of cases were antinuclear antibody positive, and 81.0% of cases were complicated with autoimmune disease. Systemic lupus erythematosus (38.1%) and Sjögren’s syndrome (23.8%) were relatively common as coexisting autoimmune diseases. Treatment was based on prednisolone use, which was used in 88.9% of patients. On the other hand, the effect of IGF-1 was limited. Overall, the prognosis of TBIR was good.

Objective To determine whether or not displaying the cost of tests can help reduce charges on test ordering in Japan. Methods This study was conducted under the setting of a simulated first visit of an outpatient for general internal medicine in a secondary medical institution in Japan. We randomly assigned 27 residents and clinical fellows to Team A or B. The first half, without charges displayed on the ordering system, was designated the "non-display group," and the participants of Team A selected tests for each paper-based simulated case (Q1-Q14), while the participants of Team B selected tests for Q15-Q28. The second half, which had charges displayed, was designated the "display group," and the participants of Team A selected tests for Q15-Q28, while the participants of Team B selected tests for Q1-Q14. The main outcome measure was the difference in the cost of tests per paper-based simulated case between the non-display and display groups. Results The median (interquartile range) cost of tests per paper-based simulated case was 12,255 yen (5,040-23,695 yen) in the non-display group versus 9,425 yen (2,320-21,700 yen) in the display group, showing a decrease of 2,830 yen with charges being displayed (p=0.002). Conclusion Displaying the charges when ordering tests in paper-based simulated cases resulted in cost reduction. The adoption of this intervention may reduce health insurance costs under the health insurance system in Japan, which has features such as universal health coverage and universal access to care.

Fumio Shimada,Yoshiyuki Ohira,Yusuke Hirota,Akiko Ikegami,Takeshi Kondo,Kiyoshi Shikino,Shingo Suzuki,Kazutaka Noda,Takanori Uehara,Masatomi Ikusaka Department of General Medicine, Chiba University Hospital, Chiba, Japan Background and objectives: Patients who come for a consultation at a general practice clinic as outpatients often suffer from background anxiety and depression. The psychological state of such patients can alleviate naturally; however, there are cases when these symptoms persist. This study investigated the realities and factors behind anxiety/depression becoming prolonged.Methods: Participants were 678 adult patients, who came to Department of General Medicine at Chiba University Hospital within a 1-year period starting from April 2012 and who completed the Hospital Anxiety and Depression Scale (HADS) during their initial consultation. Participants whose Anxiety or Depression scores in the HADS, or both, were 8 points or higher were defined as being within the anxiety/depression group, with all other participants making up the control group. A telephone interview was also conducted with participants. Furthermore, age, sex, the period from the onset of symptoms to the initial consultation at our department, the period from the initial department consultation to the telephone survey, and the existence of mental illness at the final department diagnosis were investigated.Results: A total of 121 patients (17.8% response rate) agreed to the phone survey. The HADS score during the phone survey showed that the anxiety/depression group had a significantly higher score than the control group. The HADS scores obtained between the initial consultation and telephone survey showed a positive correlation. Logistic regression analysis extracted “age” and the “continuation of the symptoms during the initial consultation” as factors that prolonged anxiety/depression.Conclusion: Anxiety and depression in general practice outpatients have the possibility of becoming prolonged for an extended period of time. Being aged 65 years or over and showing a continuation of symptoms past the initial consultation are the strongest factors associated with these prolonged conditions. When patients with anxiety and depression exhibit these risk factors, they should be further evaluated for treatment. Keywords: anxiety, depression, outpatients, general practice

Optical network failure management (ONFM) is a promising application of machine learning (ML) to optical networking. Typical ML-based ONFM approaches exploit historical monitored data, retrieved in a specific domain (e.g., a link or a network), to train supervised ML models and learn failure characteristics (a signature) that will be helpful upon future failure occurrence in that domain. Unfortunately, in operational networks, data availability often constitutes a practical limitation to the deployment of ML-based ONFM solutions, due to scarce availability of labeled data comprehensively modeling all possible failure types. One could purposely inject failures to collect training data, but this is time consuming and not desirable by operators. A possible solution is transfer learning (TL), i.e., training ML models on a source domain (SD), e.g., a laboratory testbed, and then deploying trained models on a target domain (TD), e.g., an operator network, possibly fine-tuning the learned models by re-training with few TD data. Moreover, in those cases when TL re-training is not successful (e.g., due to the intrinsic difference in SD and TD), another solution is domain adaptation, which consists of combining unlabeled SD and TD data before model training. We investigate domain adaptation and TL for failure detection and failure-cause identification across different lightpaths leveraging real optical SNR data. We find that for the considered scenarios, up to 20% points of accuracy increase can be obtained with domain adaptation for failure detection, while for failure-cause identification, only combining domain adaptation with model re-training provides significant benefit, reaching 4%–5% points of accuracy increase in the considered cases.

Objectives We examined whether problem-based learning tutorials using patient-simulated videos showing daily life are more practical for clinical learning, compared with traditional paper-based problem-based learning, for the consideration rate of psychosocial issues and the recall rate for experienced learning. Methods Twenty-two groups with 120 fifth-year students were each assigned paper-based problem-based learning and video-based problem-based learning using patient-simulated videos. We compared target achievement rates in questionnaires using the Wilcoxon signed-rank test and discussion contents diversity using the Mann-Whitney U test. A follow-up survey used a chi-square test to measure students’ recall of cases in three categories: video, paper, and non-experienced. Results Video-based problem-based learning displayed significantly higher achievement rates for imagining authentic patients (p=0.001), incorporating a comprehensive approach including psychosocial aspects (p

This randomized clinical trial compares the effect of the iExaminer teaching method on fundus examination skills with the traditional teaching method.

Network connectivity has been the traditional metric for network survivability against failures. In case of a disaster, network connectivity may not always be guaranteed due to multiple link failures. With the shifting service paradigm towards cloud computing/storage, some network services can still be provided if a content replica is available in all disconnected network segments. As a result, content connectivity has been introduced as an additional metric for network survivability under disasters. Content connectivity is defined as the reachability of content from every node in a logical topology under a given failure scenario. In this work, we investigate the content-connectivity problem in optical metro networks in the case of multiple ( $n$ ) link failures. We consider the problem of mapping a logical topology over an optical metro network such that every node in the logical topology can reach at least one data center hosting the content after $n$ -link failures. We formulate the problem as an integer linear program to minimize total network resource usage. We provide a cost comparison between content connectivity and network connectivity under typical failure scenarios.

Cases of sprue-like enteropathy associated with olmesartan have sporadically been encountered since it was first reported in 2012, and their most characteristic manifestation is severe diarrhea. We herein report the first case of sprue-like enteropathy manifesting as Wernicke-Korsakoff syndrome due to vitamin B1 malabsorption with only minimally increased bowel movements. When patients are receiving olmesartan and they complain of nonspecific chronic gastrointestinal symptoms, it is important to consider changing the drugs before any serious malabsorption syndrome develops.

In this paper, we rigorously reinvestigate the effects of childhood mortality decline during different periods on human capital accumulation and economic development. For this purpose, we extend the basic framework of Azarnert (2006) by dividing childhood among three periods—early childhood (the period prior to school enrolment), school age (the school-enrolment period), and late childhood (the period posterior to school enrolment)—and assume a constant mortality rate for each period. We arrive at the following conclusions. The mortality decline after school age promotes human capital accumulation and economic development. The mortality decline after early childhood impedes human capital accumulation and economic development, but has no effect on the same after late childhood.

A 60-year-old man was referred to us with high levels of aspartate aminotransferase and lactate dehydrogenase (LDH). He did not complain of any symptoms; however, he had been diagnosed with lumbar disc herniation, even though his back pain improved only to half of its previous level with pregabalin. Thus, we asked about the red flag of back pain and confirmed that he had involuntary body weight loss, which led us to diagnose truly non-secretory multiple myeloma, a variant of multiple myeloma that is associated with M-protein negativity and a normal serum free light chain level.

A 65-year-old Japanese man was admitted with a 4-month history of fatigue and exertional dyspnea. Transthoracic echocardiography revealed a vegetation on the aortic valve and severe aortic regurgitation. Accordingly, infective endocarditis and heart failure were diagnosed. Although a blood culture was negative on day 7 after admission, a prolonged blood culture with subculture was performed according to the patient's history of contact with cats. Consequently, Bartonella henselae was isolated. Bartonella species are fastidious bacteria that cause blood culture-negative infective endocarditis. This case demonstrates that B. henselae may be detected by prolonged incubation of blood cultures.

Brown fat dissipates energy as heat and protects against obesity. Here, we identified nuclear factor I-A (NFIA) as a transcriptional regulator of brown fat by a genome-wide open chromatin analysis of murine brown and white fat followed by motif analysis of brown-fat-specific open chromatin regions. NFIA and the master transcriptional regulator of adipogenesis, PPARγ, co-localize at the brown-fat-specific enhancers. Moreover, the binding of NFIA precedes and facilitates the binding of PPARγ, leading to increased chromatin accessibility and active transcription. Introduction of NFIA into myoblasts results in brown adipocyte differentiation. Conversely, the brown fat of NFIA knockout mice displays impaired expression of the brown-fat-specific genes and reciprocal elevation of muscle genes. Finally, expression of NFIA and the brown-fat-specific genes is positively correlated in human brown fat. These results indicate that NFIA activates the cell-type-specific enhancers and facilitates the binding of PPARγ for controlling the brown fat gene program.

Background: Human embryonic stem cell-derived cardiomyocytes (hESC-CMs) were recently shown to be capable of electromechanical integration following direct injection into intact or recently injured guinea pig hearts, and hESC-CM transplantation in recently injured hearts correlated with improvements in contractile function and a reduction in the incidence of arrhythmias. The present study was aimed at determining the ability of hESC-CMs to integrate and modulate electrical stability following transplantation in a chronic model of cardiac injury. Methods and Results: At 28 days following cardiac cryoinjury, guinea pigs underwent intracardiac injection of hESC-CMs, noncardiac hESC derivatives (non-CMs), or vehicle. Histology confirmed partial remuscularization of the infarct zone in hESC-CM recipients while non-CM recipients showed heterogeneous xenografts. The 3 experimental groups showed no significant difference in the left ventricular dimensions or fractional shortening by echocardiography or in the incidence of spontaneous arrhythmias by telemetric monitoring. Although recipients of hESC-CMs and vehicle showed a similar incidence of arrhythmias induced by programmed electrical stimulation at 4 weeks posttransplantation, non-CM recipients proved to be highly inducible, with a ∼3-fold greater incidence of induced arrhythmias. In parallel studies, we investigated the ability of hESC-CMs to couple with host myocardium in chronically injured hearts by the intravital imaging of hESC-CM grafts that stably expressed a fluorescent reporter of graft activation, the genetically encoded calcium sensor GCaMP3. In this work, we found that only ∼38% (5 of 13) of recipients of GCaMP3+ hESC-CMs showed fluorescent transients that were coupled to the host electrocardiogram. Conclusions: Human embryonic stem cell-derived cardiomyocytes engraft in chronically injured hearts without increasing the incidence of arrhythmias, but their electromechanical integration is more limited than previously reported following their transplantation in a subacute injury model. Moreover, non-CM grafts may promote arrhythmias under certain conditions, a finding that underscores the need for input preparations of high cardiac purity.

A 65-year-old Japanese man was admitted with a 4-month history of fatigue and exertional dyspnea. Transthoracic echocardiography revealed a vegetation on the aortic valve and severe aortic regurgitation. Accordingly, infective endocarditis and heart failure were diagnosed. Although a blood culture was negative on day 7 after admission, a prolonged blood culture with subculture was performed according to the patient's history of contact with cats. Consequently, Bartonella henselae was isolated. Bartonella species are fastidious bacteria that cause blood culture-negative infective endocarditis. This case demonstrates that B. henselae may be detected by prolonged incubation of blood cultures.

Learning point for clinicians Acute sphenoiditis may affect the second branch of the trigeminal nerve because of anatomical reasons and should be included in differential diagnoses in patients with headache accompanied by facial numbness. A 67-year-old man visited our hospital with complaints of headache, vomiting and numbness on the right side of the face. He noted preceding transient upper respiratory tract symptoms 7 days before his visit. The headache that occurred 2 days before his visit was the most severe he had ever experienced and worsened. On the day before his visit, it was accompanied by vomiting and numbness on the right side of the face. Physical examination showed a temperature of 36.6°C, blood pressure of 167/83 mmHg, pulse rate of 72 beats per minute, and oxygen saturation as measured using pulse oximetry (SpO2) of …

Symptoms of rhabdomyolysis: myalgia, muscle weakness, dark urine, and, in some, muscle swelling.

The Kobe (CK4) meteorite and other CK chondrites are characterized with respect to rare earth element (REE) abundances using 12 bulk samples from 10 CK chondrites including Ningqiang (CK3), Karoonda (CK4), Kobe (CK4), Y-693 (CK4), A-882113 (CK4), Maralinga (CK4), Y-82102 (CK5), Y-82105 (CK5), Y-82191 (CK6) and A-881551 (CK6). REE, Ba, Sr, Ca, Mg, Fe, Rb and K were analyzed by high-precision isotope dilution. The CK chondrites examined exhibit systematically higher REE abundances compared to ordinary chondrites, comparable to CV and CO, flat REE patterns with minor negative anomalies up to -15% on Ce and -25% on Eu. These chondrites also exhibit an appreciable light/heavy REE gap and Yb anomaly compared to Cl chondrites. CK3-4 chondrites in general are found to exhibit a larger negative Ce anomaly (mean anomaly -9.6 ′3.8%) compared to CK5-6 chondrites (3.3 ′ 1.0%), and some CK5-6 show a slightly light-REE depleted pattern with less-pronounced negative Ce anomaly. It is suggested that these CK chondrites had a larger negative Ce anomaly prior to early thermal metamorphism, which yielded the minor redistribution of light REEs. No clear correlation is found between Sm and Eu/Sm or between Eu and Sr. From these observations, it is suggested that Eu existed in the trivalent state due to higher oxidizing conditions, suppressing redistribution of REEs among minerals compared to metamorphosed ordinary chondrites during the early thermal metamorphic event on the parent body. The systematic REE anomalies observed in CK chondrites are considered to be inherited properties from the refractory precursors produced in the early solar nebula.

A 75-year-old man presented with a 2-month history of worsening chest and back pain. Physical examination revealed generalised bone tenderness with associated indirect pain; digital rectal examination indicated asymmetric prostatic induration. Laboratory examination revealed high serum alkaline phosphatase (1348 U/L (bone type, 73.2%)), high C reactive protein (4.5 mg/dL), high serum ferritin (2952 ng/mL) and high prostate-specific antigen (1346 ng/mL). A whole-body technetium-99m methyldiphosphonate (Tc-99m MDP) scan revealed increase in bone …

A 55-year-old male had been experiencing pain over the left shoulder to the anterior part of the chest since two days. Physical examination indicated a body temperature of 37.8°C, as well as swelling, elevated temperature, and tenderness on the left side of the sternum (Figure 1). The pain worsened with elevation of the left shoulder, right rotation and retroflexion of the neck, and deep inspiration. Blood tests showed an inflammatory response (C-reactive protein, 13.1 mg/dL, erythrocyte sedimentation rate, 48 mm/h). Rheumatoid factor and anti-cyclic citrullinated peptide (anti-CCP) antibody were negative. Joint aspiration was not performed because ultrasonography did not detect the effusion of sternoclavicular joints. Computed tomography (CT) scan of the sternoclavicular joints revealed an erosive change in the left sternoclavicular joint. Magnetic resonance with the short-T1 inversion-recovery (MR-STIR) sequence showed an abnormal signal in the left sternoclavicular joint (Figure 2). A non-steroidal anti-inflammatory drug was orally administered, and the symptoms disappeared after one month. Although no skin eruption developed during the follow-up, sternoclavicular joint arthritis due to SAPHO syndrome was diagnosed.

The purpose of this study is to develop a system analyzing cell activity by the dielectrophoresis method. Our previous studies revealed a correlation between the growth activity and dielectric property (Re[K(ω)]) of mouse hybridoma 3-2H3 cells using dielectrophoretic levitation. Furthermore, it was clarified that the differentiation activity of many stem cells could be evaluated by the Re[K(ω)] without differentiation induction. In this paper, 3-2H3 cells exposed to an alternating current (AC) electric field or a direct current (DC) electric field were cultivated, and the influence of damage by the electric field on the growth activity of the cells was examined. To evaluate the activity of the cells by measuring the Re[K(ω)], the correlation between the growth activity and the Re[K(ω)] of the cells exposed to the electric field was examined. The relations between the cell viability, growth activity, and Re[K(ω)] in the cells exposed to the AC electric field were obtained. The growth activity of the cells exposed to the AC electric field could be evaluated by the Re[K(ω)]. Furthermore, it was found that the adverse effects of the electric field on the cell viability and the growth activity were smaller in the AC electric field than the DC electric field.

We develop an SDN-based control for optical-multicast packet transmission and experimentally demonstrate multicast functionality by validating it using an application-layer network service for efficient content duplication in Optical Packet/Circuit Integrated (OPCI) network.