The monitoring of oil production process by deep learning based on morphology in oleaginous yeasts.

BACKGROUND: Monitoring jar fermenter-cultured microorganisms in real time is important for controlling productivity of bioproducts in large-scale cultivation settings. Morphological data is used to understand the growth and fermentation states of these microorganisms during monitoring. Oleaginous yeasts are used for their high productivity of single-cell oils but the relationship between lipid productivity and morphology has not been elucidated in these organisms. RESULTS: In this study, we investigated the relationship between the morphology of oleaginous yeasts (Lipomyces starkeyi and Rhodosporidium toruloides were used) and their cultivation state in a large-scale cultivation setting using a real-time monitoring system. We combined this with deep learning by feeding a large amount of high-definition cell images obtained from the monitoring system to a deep learning algorithm. Our results showed that the cell images could be grouped into 7 distinct groups and that a strong correlation existed between each group and its biochemical activity (growth and oil-productivity). CONCLUSIONS: This is the first report describing the morphological variations of oleaginous yeasts in a large-scale cultivation, and describes a promising new avenue for improving productivity of microorganisms in large-scale cultivation through the use of a real-time monitoring system combined with deep learning. KEY POINTS: • A real-time monitoring system followed the morphological change of oleaginous yeasts. • Deep learning grouped them into 7 distinct groups based on their morphology. • A correlation between the cultivation state and the shape of the yeast was observed.

Automatic detection of fish sounds based on multi-stage classification including logistic regression via adaptive feature weighting.

This paper presents a method for automatic detection of fish sounds in an underwater environment. There exist two difficulties: (i) features and classifiers that provide good detection results differ depending on the underwater environment and (ii) there are cases where a large amount of training data that is necessary for supervised machine learning cannot be prepared. A method presented in this paper (the proposed hybrid method) overcomes these difficulties as follows. First, novel logistic regression (NLR) is derived via adaptive feature weighting by focusing on the accuracy of classification results by multiple classifiers, support vector machine (SVM), and k-nearest neighbors (k-NN). Although there are cases where SVM or k-NN cannot work well due to divergence of useful features, NLR can produce complementary results. Second, the proposed hybrid method performs multi-stage classification with consideration of the accuracy of SVM, k-NN, and NLR. The multi-stage acquisition of reliable results works adaptively according to the underwater environment to reduce performance degradation due to diversity of useful classifiers even if abundant training data cannot be prepared. Experiments on underwater recordings including sounds of Sciaenidae such as silver croakers (Pennahia argentata) and blue drums (Nibea mitsukurii) show the effectiveness of the proposed hybrid method.

Trend clustering from COVID-19 tweets using graphical lasso-guided iterative principal component analysis.

This article presents a method for trend clustering from tweets about coronavirus disease (COVID-19) to help us objectively review the past and make decisions about future countermeasures. We aim to avoid detecting usual trends based on seasonal events while detecting essential trends caused by the influence of COVID-19. To this aim, we regard daily changes in the frequencies of each word in tweets as time series signals and define time series signals with single peaks as target trends. To successfully cluster the target trends, we propose graphical lasso-guided iterative principal component analysis (GLIPCA). GLIPCA enables us to remove trends with indirect correlations generated by other essential trends. Moreover, GLIPCA overcomes the difficulty in the quantitative evaluation of the accuracy of trend clustering. Thus, GLIPCA's parameters are easier to determine than those of other clustering methods. We conducted experiments using Japanese tweets about COVID-19 from March 8, 2020, to May 7, 2020. The results show that GLIPCA successfully distinguished trends before and after the declaration of a state of emergency on April 7, 2020. In addition, the results reveal the international argument about whether the Tokyo 2020 Summer Olympics should be held. The results suggest the tremendous social impact of the words and actions of Japanese celebrities. Furthermore, the results suggest that people's attention moved from worry and fear of an unknown novel pneumonia to the need for medical care and a new lifestyle as well as the scientific characteristics of COVID-19.

Ranking of Importance Measures of Tweet Communities: Application to Keyword Extraction From COVID-19 Tweets in Japan.

This article presents a method that detects tweet communities with similar topics and ranks the communities by importance measures. By identifying the tweet communities that have high importance measures, it is possible for users to easily find important information about the coronavirus disease (COVID-19). Specifically, we first construct a community network, whose nodes are tweet communities obtained by applying a community detection method to a tweet network. The community network is constructed based on textual similarities between tweet communities and sizes of tweet communities. Second, we apply algorithms for calculating centrality to the community network. Because the obtained centrality is based on tweet community sizes as well, we call it the importance measure in distinction to conventional centrality. The importance measure can simultaneously evaluate the importance of topics in the entire data set and occupancy (or dominance) of tweet communities in the network structure. We conducted experiments by collecting Japanese tweets about COVID-19 from March 1, 2020 to May 15, 2020. The results show that the proposed method is able to extract keywords that have a high correlation with the number of people infected with COVID-19 in Japan. Because users can browse the keywords from a small number of central tweet communities, quick and easy understanding of important information becomes feasible.

River state classification combining patch-based processing and CNN.

This paper proposes a method for classifying the river state (a flood risk exists or not) from river surveillance camera images by combining patch-based processing and a convolutional neural network (CNN). Although CNN needs much training data, the number of river surveillance camera images is limited because flood does not frequently occur. Also, river surveillance camera images include objects that are irrelevant to the flood risk. Therefore, the direct use of CNN may not work well for the river state classification. To overcome this limitation, this paper develops patch-based processing for adjusting CNN to the river state classification. By increasing training data via the patch segmentation of an image and selecting patches that are relevant to the river state, the adjustment of general CNNs to the river state classification becomes feasible. The proposed patch-based processing and CNN are developed independently. This yields the practical merits that any CNN can be used according to each user's purposes, and the maintenance and improvement of each component of the whole system can be easily performed. In the experiment, river state classification is defined as the following problems using two datasets, to verify the effectiveness of the proposed method. First, river images from the public dataset called Places are classified to images with Muddy labels and images with Clear labels. Second, images from the river surveillance camera in Nagaoka City, Japan are classified to images captured when the government announced heavy rain or flood warning and the other images.

The three-axial gyroscope sensor detects the turning point between opening and closing phases of chewing.

Most devices measuring the kinematics of masticatory function are cumbersome to setup and not portable. Data collection would be facilitated, particularly in the elderly, if the device used for the objective measurement of mastication was easily transportable and simple to setup. Accelerometers and gyroscope sensors are lightweight and portable and may be useful alternatives. The definition of the turning point between the opening and closing phases of chewing is important for studies of associations between muscle activity and effects of perturbations. Measures of the mediolateral angle (specifically, the mandibular tilt from the lateral view) allow the detection of the turning point between the opening and closing phases. The aim was to determine whether a three-axial gyroscope sensor can detect the turning point between opening and closing phases of chewing. Fourteen asymptomatic participants chewed gum while the output was recorded from a three-axial gyroscope sensor (Seiko Epson, Japan) attached to the chin and a 6 degree-of-freedom electromagnetic jaw-tracking device (Pollhemus, USA). Bland-Altman plots were used to assess the matching of the recordings made by the three-axial gyroscope sensor and the jaw-tracking device. The turning points between the opening and closing phases of chewing matched closely when recorded by a jaw-tracking device and when using a three-axial gyroscope sensor. A three-axial gyroscope sensor can validly detect the turning point between the opening and closing phases during chewing of gum.