Language：English   Publisher：Neural Computation  

Generative modeling of crystal structures is significantly challenged by the complexity of input data, which constrains the ability of these models to explore and discover novel crystals. This complexity often confines de novo design methodologies to merely small perturbations of known crystals and hampers the effective application of advanced optimization techniques. One such optimization technique, latent space Bayesian optimization (LSBO), has demonstrated promising results in uncovering novel objects across various domains, especially when combined with variational autoencoders (VAEs). Recognizing LSBO’s potential and the critical need for innovative crystal discovery, we introduce Crystal-LSBO, a de novo design framework for crystals specifically tailored to enhance explorability within LSBO frameworks. Crystal-LSBO employs multiple VAEs, each dedicated to a distinct aspect of crystal structure—lattice, coordinates, and chemical elements—orchestrated by an integrative model that synthesizes these components into a cohesive output. This setup not only streamlines the learning process but also produces explorable latent spaces thanks to the decreased complexity of the learning task for each model, enabling LSBO approaches to operate. Our study pioneers the use of LSBO for de novo crystal design, demonstrating its efficacy through optimization tasks focused mainly on formation energy values. Our results highlight the effectiveness of our methodology, offering a new perspective for de novo crystal discovery.

DOI： 10.1162/neco_a_01767

Open Access

Web of Science

Scopus

PubMed

Publisher：Digital Discovery  

Self-driving laboratories (SDLs) are transforming the scientific discovery process worldwide by integrating automated experimentation with data-driven decision-making. Japan, known for its automation industry, is actively contributing to this field. This perspective introduces Japan's efforts in SDL development, including diverse applications across materials science, biology, chemistry, and software. In addition, it covers national funding programs, research communities, and Japanese industries supporting progress in this field. It also highlights the importance of education, standardization, and benchmarking for the future growth of SDL research.

DOI： 10.1039/d4dd00387j

Open Access

Web of Science

Scopus

Authorship：Last author,　Corresponding author  

Authorship：Last author,　Corresponding author   Language：English   Publishing type：Research paper (international conference proceedings)  

Language：English   Publishing type：Research paper (international conference proceedings)  

Language：English   Publishing type：Research paper (international conference proceedings)  

Language：English   Publishing type：Research paper (scientific journal)  

DOI： 10.1162/neco_a_01724

Authorship：Lead author,　Last author,　Corresponding author   Language：English   Publishing type：Research paper (scientific journal)  

DOI： 10.1002/tee.24262

Open Access

Publisher：Proceedings of the ACM SIGKDD International Conference on Knowledge Discovery and Data Mining  

Selecting the optimal combination of a machine learning (ML) algorithm and its hyper-parameters is crucial for the development of high-performance ML systems. However, since the combination of ML algorithms and hyper-parameters is enormous, the exhaustive validation requires a significant amount of time. Many existing studies use Bayesian optimization (BO) for accelerating the search. On the other hand, a significant difficulty is that, in general, there exists a different hyper-parameter space for each one of candidate ML algorithms. BO-based approaches typically build a surrogate model independently for each hyper-parameter space, by which sufficient observations are required for all candidate ML algorithms. In this study, our proposed method embeds different hyper-parameter spaces into a shared latent space, in which a surrogate multi-task model for BO is estimated. This approach can share information of observations from different ML algorithms by which efficient optimization is expected with a smaller number of total observations. We further propose the pre-training of the latent space embedding with an adversarial regularization, and a ranking model for selecting an effective pre-trained embedding for a given target dataset. Our empirical study demonstrates effectiveness of the proposed method through datasets from OpenML.

DOI： 10.1145/3711896.3736851

Web of Science

Scopus

Publisher：Transactions on Machine Learning Research  

Bayesian optimization is a powerful tool for optimizing an expensive-to-evaluate black-box function. In particular, the effectiveness of expected improvement (EI) has been demonstrated in a wide range of applications. However, theoretical analyses of EI are limited compared with other theoretically established algorithms. This paper analyzes a randomized variant of EI, which evaluates the EI from the maximum of the posterior sample path. We show that this posterior sampling-based random EI achieves the sublinear Bayesian cumulative regret bounds under the assumption that the black-box function follows a Gaussian process. Finally, we demonstrate the effectiveness of the proposed method through numerical experiments.

Scopus

Publisher：Transactions on Machine Learning Research  

Coreset selection, which involves selecting a small subset from an existing training dataset, is an approach to reducing training data size, and various approaches have been proposed for this method. In practical situations where these methods are employed, it is often the case that the data distributions differ between the development phase and the deployment phase, with the latter being unknown. Thus, it is challenging to select an effective subset of training data that performs well across all deployment scenarios. We therefore propose Distributionally Robust Coreset Selection (DRCS), which theoretically derives an estimate of the upper bound for the worst-case test error, assuming that the future covariate distribution may deviate within a defined range from the training distribution. Furthermore, by selecting instances in a way that suppresses the estimate of the upper bound for the worst-case test error, DRCS achieves distributionally robust training instance selection. This study is primarily applicable to convex training computation, but we demonstrate that it can also be applied to deep learning under appropriate approximations. In this paper, we focus on covariate shift, a type of data distribution shift, and demonstrate the effectiveness of DRCS through experiments.

Scopus

Publisher：Proceedings of Machine Learning Research  

Gaussian process regression (GPR) or kernel ridge regression is a widely used and powerful tool for nonlinear prediction. Therefore, active learning (AL) for GPR, which actively collects data labels to achieve an accurate prediction with fewer data labels, is an important problem. However, existing AL methods do not theoretically guarantee prediction accuracy for target distribution. Furthermore, as discussed in the distributionally robust learning literature, specifying the target distribution is often difficult. Thus, this paper proposes two AL methods that effectively reduce the worst-case expected error for GPR, which is the worst-case expectation in target distribution candidates. We show an upper bound of the worst-case expected squared error, which suggests that the error will be arbitrarily small by a finite number of data labels under mild conditions. Finally, we demonstrate the effectiveness of the proposed methods through synthetic and real-world datasets.

Scopus

Publisher：Transactions on Machine Learning Research  

The rapid discovery of new chemical compounds is essential for advancing global health and developing treatments. While generative models show promise in creating novel molecules, challenges remain in ensuring the real-world applicability of these molecules and finding such molecules efficiently. To address this challenge, we introduce Conditional Latent Space Molecular Scaffold Optimization (CLaSMO), which integrates a Conditional Variational Autoencoder (CVAE) with Latent Space Bayesian Optimization (LSBO) to strategically modify molecules while preserving similarity to the original input, effectively framing the task as constrained optimization. Our LSBO setting improves the sample-efficiency of the molecular optimization, and our modification approach helps us to obtain molecules with higher chances of real-world applicability. CLaSMO explores substructures of molecules in a sample-efficient manner by performing BO in the latent space of a CVAE conditioned on the atomic environment of the molecule to be optimized. Our extensive evaluations across diverse optimization tasks—including rediscovery, docking score, and multi-property opti-mization—show that CLaSMO efficiently enhances target properties, delivers remarkable sample-efficiency crucial for resource-limited applications while considering molecular similarity constraints, achieves state of the art performance, and maintains practical synthetic accessibility. We also provide an open-source web application¹ that enables chemical experts to apply CLaSMO in a Human-in-the-Loop setting.

Scopus

Publisher：Transactions on Machine Learning Research  

Effective condition monitoring in complex systems requires identifying change points (CPs) in the frequency domain, as the structural changes often arise across multiple frequencies. This paper extends recent advancements in statistically significant CP detection, based on Selective Inference (SI), to the frequency domain. The proposed SI method quantifies the statistical significance of detected CPs in the frequency domain using p-values, ensuring that the detected changes reflect genuine structural shifts in the target system. We address two major technical challenges to achieve this. First, we extend the existing SI framework to the frequency domain by appropriately utilizing the properties of discrete Fourier transform (DFT). Second, we develop an SI method that provides valid p-values for CPs where changes occur across multiple frequencies. Experimental results demonstrate that the proposed method reliably identifies genuine CPs with strong statistical guarantees, enabling more accurate root-cause analysis in the frequency domain of complex systems.

Scopus

Publisher：Proceedings of Machine Learning Research  

A data analysis pipeline is a structured sequence of steps that transforms raw data into meaningful insights by integrating various analysis algorithms. In this paper, we propose a novel statistical test to assess the significance of data analysis pipelines. Our approach enables the systematic development of valid statistical tests applicable to any feature selection pipeline composed of predefined components. We develop this framework based on selective inference, a statistical technique that has recently gained attention for data-driven hypotheses. As a proof of concept, we focus on feature selection pipelines for linear models, composed of three missing value imputation algorithms, three outlier detection algorithms, and three feature selection algorithms. We theoretically prove that our statistical test can control the probability of false positive feature selection at any desired level, and demonstrate its validity and effectiveness through experiments on synthetic and real data. Additionally, we present an implementation framework that facilitates testing across any configuration of these feature selection pipelines without extra implementation costs.

Scopus

Publisher：Transactions on Machine Learning Research  

Graph Neural Networks (GNNs) have gained prominence for their ability to process graph-structured data across various domains. However, interpreting GNN decisions remains a significant challenge, leading to the adoption of saliency maps for identifying salient sub-graphs composed of influential nodes and edges. Despite their utility, the reliability of GNN saliency maps has been questioned, particularly in terms of their robustness to input noise. In this study, we propose a statistical testing framework to rigorously evaluate the significance of saliency maps. Our main contribution lies in addressing the inflation of the Type I error rate caused by double-dipping of data, leveraging the framework of Selective Inference. Our method provides statistically valid p-values while controlling the Type I error rate, en-suring that identified salient subgraphs contain meaningful information rather than random artifacts. The method is applicable to a variety of saliency methods with piecewise linear-ity (e.g., Class Activation Mapping). We validate our method on synthetic and real-world datasets, demonstrating its capability in assessing the reliability of GNN interpretations.

Scopus

DOI： 10.1080/27660400.2024.2336402

Open Access

Web of Science

Language：English   Publisher：Neural Computation  

In this study, we investigate the quantification of the statistical reliability of detected change points (CPs) in time series using a recurrent neural network (RNN). Thanks to its flexibility, RNN holds the potential to effectively identify CPs in time series characterized by complex dynamics. However, there is an increased risk of erroneously detecting random noise fluctuations as CPs. The primary goal of this study is to rigorously control the risk of false detections by providing theoretically valid (Formula presented.) -values to the CPs detected by RNN. To achieve this, we introduce a novel method based on the framework of selective inference (SI). SI enables valid inferences by conditioning on the event of hypothesis selection, thus mitigating bias from generating and testing hypotheses on the same data. In this study, we apply an SI framework to RNN-based CP detection, where characterizing the complex process of RNN selecting CPs is our main technical challenge. We demonstrate the validity and effectiveness of the proposed method through artificial and real data experiments.

DOI： 10.1162/neco_a_01724

Web of Science

Scopus

PubMed

Language：English   Publisher：Journal of Pathology Informatics  

In this study, we present a deep-learning-based multimodal classification method for lymphoma diagnosis in digital pathology, which utilizes a whole slide image (WSI) as the primary image data and flow cytometry (FCM) data as auxiliary information. In pathological diagnosis of malignant lymphoma, FCM serves as valuable auxiliary information during the diagnosis process, offering useful insights into predicting the major class (superclass) of subtypes. By incorporating both images and FCM data into the classification process, we can develop a method that mimics the diagnostic process of pathologists, enhancing the explainability. In order to incorporate the hierarchical structure between superclasses and their subclasses, the proposed method utilizes a network structure that effectively combines the mixture of experts (MoE) and multiple instance learning (MIL) techniques, where MIL is widely recognized for its effectiveness in handling WSIs in digital pathology. The MoE network in the proposed method consists of a gating network for superclass classification and multiple expert networks for (sub)class classification, specialized for each superclass. To evaluate the effectiveness of our method, we conducted experiments involving a six-class classification task using 600 lymphoma cases. The proposed method achieved a classification accuracy of 72.3%, surpassing the 69.5% obtained through the straightforward combination of FCM and images, as well as the 70.2% achieved by the method using only images. Moreover, the combination of multiple weights in the MoE and MIL allows for the visualization of specific cellular and tumor regions, resulting in a highly explanatory model that cannot be attained with conventional methods. It is anticipated that by targeting a larger number of classes and increasing the number of expert networks, the proposed method could be effectively applied to the real problem of lymphoma diagnosis.

DOI： 10.1016/j.jpi.2023.100359

Open Access

Scopus

PubMed

Publisher：Japanese Journal of Statistics and Data Science  

Selective inference (SI) has been actively studied as a promising framework for statistical hypothesis testing for data-driven hypotheses. The basic idea of SI is to make inferences conditional on an event that a hypothesis is selected. To perform SI, this event must be characterized in a traceable form. When selection event is too difficult to characterize, additional conditions are introduced for tractability. This additional conditions often cause the loss of power, and this issue is referred to as over-conditioning in Fithian et al. (Optimal inference after model selection, arXiv preprint arXiv:1410.2597, 2014). Parametric programming-based SI (PP-based SI) has been proposed as one way to address the over-conditioning issue. The main problem of PP-based SI is its high computational cost due to the need to exhaustively explore the data space. In this study, we introduce a procedure to reduce the computational cost while guaranteeing the desired precision, by proposing a method to compute the lower and upper bounds of p values. We also proposed three types of search strategies that efficiently improve these bounds. We demonstrate the effectiveness of the proposed method in hypothesis testing problems for feature selection in linear models and attention region identification in deep neural networks.

DOI： 10.1007/s42081-024-00247-0

Web of Science

Scopus

Language：English   Publisher：Neural Computation  

Latent space Bayesian optimization (LSBO) combines generative models, typically variational autoencoders (VAE), with Bayesian optimization (BO), to generate de novo objects of interest. However, LSBO faces challenges due to the mismatch between the objectives of BO and VAE, resulting in poor exploration capabilities. In this article, we propose novel contributions to enhance LSBO efficiency and overcome this challenge. We first introduce the concept of latent consistency/inconsistency as a crucial problem in LSBO, arising from the VAE-BO mismatch. To address this, we propose the latent consistent aware-acquisition function (LCA-AF) that leverages consistent points in LSBO. Additionally, we present LCA-VAE, a novel VAE method that creates a latent space with increased consistent points through data augmentation in latent space and penalization of latent inconsistencies. Combining LCA-VAE and LCA-AF, we develop LCA-LSBO. Our approach achieves high sample efficiency and effective exploration, emphasizing the significance of addressing latent consistency through the novel incorporation of data augmentation in latent space within LCA-VAE in LSBO. We showcase the performance of our proposal via de novo image generation and de novo chemical design tasks.

DOI： 10.1162/neco_a_01708

Web of Science

Scopus

PubMed

Language：English   Publishing type：Research paper (scientific journal)  

DOI： 10.1515/jqas-2023-0039

Authorship：Last author,　Corresponding author   Language：English   Publishing type：Research paper (international conference proceedings)  

DOI： 10.5555/3692070.3693905

Language：English   Publishing type：Research paper (international conference proceedings)  

DOI： 10.5555/3692070.3694005

Authorship：Last author,　Corresponding author   Language：English   Publishing type：Research paper (international conference proceedings)  

Open Access

Authorship：Last author   Language：English   Publishing type：Research paper (international conference proceedings)  

Open Access

Language：English   Publishing type：Research paper (scientific journal)  

DOI： 10.1016/j.micron.2024.103663

Open Access

Authorship：Last author,　Corresponding author   Language：English   Publishing type：Research paper (scientific journal)  

DOI： 10.1007/s42081-024-00247-0

Language：English   Publishing type：Research paper (scientific journal)  

Language：English   Publisher：Laboratory Investigation  

Pathologic evaluation is the most crucial method for diagnosing malignant lymphomas. However, there are no established diagnostic criteria for evaluating pathologic morphology. We manually circled cell nuclei in the lesions of 10 patients with diffuse large B-cell lymphoma (DLBCL), follicular lymphoma, and reactive lymphadenitis. Seventeen parameters related to nuclear shape, color, and other characteristics were measured. We attempted to compare the statistical differences between these subtypes and extract distinctive disease-specific populations on the basis of these parameters. Statistically significant differences were observed between the different types of lymphoma for many of the 17 parameters. Through t-distributed stochastic neighbor embedding analysis, we extracted a cluster of cells that showed distinctive features of DLBCL and were not found in follicular lymphoma or reactive lymphadenitis. We created a decision tree to identify the characteristics of the cells within that cluster. Based on a 5-fold cross-validation study, the average sensitivity, specificity, and accuracy obtained were 84.1%, 98.4%, and 97.3%, respectively. A similar result was achieved using a validation experiment. Important parameters that indicate the features of DLBCL include Area, ConcaveCount, MaxGray, and ModeGray. By quantifying pathologic morphology, it was possible to objectively represent the cell morphology specific to each lymphoma subtype using quantitative indicators. The quantified morphologic information has the potential to serve as a reproducible and flexible diagnostic tool.

DOI： 10.1016/j.labinv.2023.100302

Web of Science

Scopus

PubMed

Publisher：Stat  

In predictive modelling for high-stake decision-making, predictors must be not only accurate but also reliable. Conformal prediction (CP) is a promising approach for obtaining the coverage of prediction results with fewer theoretical assumptions. To obtain the prediction set by so-called full-CP, we need to refit the predictor for all possible values of prediction results, which is only possible for simple predictors. For complex predictors such as random forests (RFs) or neural networks (NNs), split-CP is often employed where the data is split into two parts: one part for fitting and another for computing the prediction set. Unfortunately, because of the reduced sample size, split-CP is inferior to full-CP both in fitting as well as prediction set computation. In this paper, we develop a full-CP of sparse high-order interaction model (SHIM), which is sufficiently flexible as it can take into account high-order interactions among variables. We resolve the computational challenge for full-CP of SHIM by introducing a novel approach called homotopy mining. Through numerical experiments, we demonstrate that SHIM is as accurate as complex predictors such as RF and NN and enjoys the superior statistical power of full-CP.

DOI： 10.1002/sta4.633

Open Access

Web of Science

Scopus

Publisher：Transactions on Machine Learning Research  

Level set estimation (LSE) the problem of identifying the set of input points where a function takes a value above (or below) a given threshold is important in practical applications. When the function is expensive to evaluate and black-box, the straddle algorithm, a representative heuristic for LSE based on Gaussian process models, and its extensions with theoretical guarantees have been developed. However, many existing methods include a confidence parameter, β^1/2 t, that must be specified by the user. Methods that choose β^1/2 t heuristically do not provide theoretical guarantees. In contrast, theoretically guaranteed values of β^1/2 t need to be increased depending on the number of iterations and candidate points; they are conservative and do not perform well in practice. In this study, we propose a novel method, the randomized straddle algorithm, in which β<inf>t</inf> in the straddle algorithm is replaced by a random sample from the chi-squared distribution with two degrees of freedom. The confidence parameter in the proposed method does not require adjustment, does not depend on the number of iterations and candidate points, and is not conservative. Furthermore, we show that the proposed method has theoretical guarantees that depend on the sample complexity and the number of iterations. Finally, we validate the applicability of the proposed method through numerical experiments using synthetic and real data.

Scopus

Publisher：Proceedings of the International Joint Conference on Computer Vision Imaging and Computer Graphics Theory and Applications  

We propose a method that modifies encoding in DDIM (Denoising Diffusion Implicit Model) to improve the quality of counterfactual histopathological images of malignant lymphoma. Counterfactual medical images are widely employed for analyzing the changes in images accompanying disease. For the analysis of pathological images, it is desired to accurately represent the types of individual cells in the tissue. We employ DDIM because it can refer to exogenous variables in causal models and can generate counterfactual images. Here, one problem of DDIM is that it does not always generate accurate images due to approximations in the forward process. In this paper, we propose a method that reduces the errors in the encoded images obtained in the forward process. Since the computation in the backward process of DDIM does not include any approximation, the accurate encoding in the forward process can improve the accuracy of the image generation. Our proposed method improves the accuracy of encoding by explicitly referring to the given original image. Experiments demonstrate that our proposed method accurately reconstructs original images, including microstructures such as cell nuclei, and outperforms the conventional DDIM in several measures of image generation.

DOI： 10.5220/0012366100003660

Open Access

Scopus

Publisher：Lecture Notes in Computer Science Including Subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics  

Whole-slide image analysis via the means of computational pathology often relies on processing tessellated gigapixel images with only slide-level labels available. Applying multiple instance learning-based methods or transformer models is computationally expensive as, for each image, all instances have to be processed simultaneously. The MLP-Mixer is an under-explored alternative model to common vision transformers, especially for large-scale datasets. Due to the lack of a self-attention mechanism, they have linear computational complexity to the number of input patches but achieve comparable performance on natural image datasets. We propose a combination of feature embedding and clustering to preprocess the full whole-slide image into a reduced prototype representation which can then serve as input to a suitable MLP-Mixer architecture. Our experiments on two public benchmarks and one inhouse malignant lymphoma dataset show comparable performance to current state-of-the-art methods, while achieving lower training costs in terms of computational time and memory load. Code is publicly available at https://github.com/butkej/ProtoMixer.

DOI： 10.1007/978-3-031-45676-3_12

Web of Science

Scopus

Language：English   Publisher：The Japan Society of Mechanical Engineers  

<p>This study employs artificial neural networks (ANNs) to predict the structural compliance of randomly generated irregular geometries derived from Finite Element (FE) calculations. By imposing volume constraints, the scope of the study is confined to applying ANNs for learning from structural data generated by considering either multiple random walks of a circle or a set of randomly placed circles with allowed overlaps. Numerical results indicate that the learning outcomes of the former approach are more satisfactory than those of the latter. This suggests that the effectiveness of employing ANNs for predicting the structural compliance of irregular geometries is contingent upon how the random geometries are generated and the material volume ratio. The learning outcomes of irregular structures generated by the former approach with a higher volume ratio exhibit greater satisfaction due to a higher degree of structural connectivity.</p>

DOI： 10.1299/mej.24-00002

Open Access

Web of Science

CiNii Research

Language：English   Publisher：Patterns  

Predictive pattern mining is an approach used to construct prediction models when the input is represented by structured data, such as sets, graphs, and sequences. The main idea behind predictive pattern mining is to build a prediction model by considering unified inconsistent notation sub-structures, such as subsets, subgraphs, and subsequences (referred to as patterns), present in the structured data as features of the model. The primary challenge in predictive pattern mining lies in the exponential growth of the number of patterns with the complexity of the structured data. In this study, we propose the safe pattern pruning method to address the explosion of pattern numbers in predictive pattern mining. We also discuss how it can be effectively employed throughout the entire model building process in practical data analysis. To demonstrate the effectiveness of the proposed method, we conduct numerical experiments on regression and classification problems involving sets, graphs, and sequences.

DOI： 10.1016/j.patter.2023.100890

Open Access

Web of Science

Scopus

PubMed

Language：English   Publisher：Neural Computation  

In this study, we have developed an incremental machine learning (ML) method that efficiently obtains the optimal model when a small number of instances or features are added or removed. This problem holds prac-tical importance in model selection, such as cross-validation (CV) and feature selection. Among the class of ML methods known as linear es-timators, there exists an efficient model update framework, the low-rank update, that can effectively handle changes in a small number of rows and columns within the data matrix. However, for ML methods beyond linear estimators, there is currently no comprehensive framework available to obtain knowledge about the updated solution within a specific computational complexity. In light of this, our study introduces a the generalized low-rank update (GLRU) method, which extends the low-rank update framework of linear estimators to ML methods formulated as a certain class of regularized empirical risk minimization, including commonly used methods such as support vector machines and logistic regression. The proposed GLRU method not only expands the range of its applicabil-ity but also provides information about the updated solutions with a computational complexity proportional to the number of data set changes. To demonstrate the effectiveness of the GLRU method, we conduct ex-periments showcasing its efficiency in performing cross-validation and feature selection compared to other baseline methods.

DOI： 10.1162/neco_a_01619

Web of Science

Scopus

PubMed

Language：English   Publisher：Neurosurgical Review  

Accurate tumor identification during surgical excision is necessary for neurosurgeons to determine the extent of resection without damaging the surrounding tissues. No conventional technologies have achieved reliable performance for pituitary adenomas. This study proposes a deep learning approach using intraoperative endoscopic images to discriminate pituitary adenomas from non-tumorous tissue inside the sella turcica. Static images were extracted from 50 intraoperative videos of patients with pituitary adenomas. All patients underwent endoscopic transsphenoidal surgery with a 4 K ultrahigh-definition endoscope. The tumor and non-tumorous tissue within the sella turcica were delineated on static images. Using intraoperative images, we developed and validated deep learning models to identify tumorous tissue. Model performance was evaluated using a fivefold per-patient methodology. As a proof-of-concept, the model’s predictions were pathologically cross-referenced with a medical professional’s diagnosis using the intraoperative images of a prospectively enrolled patient. In total, 605 static images were obtained. Among the cropped 117,223 patches, 58,088 were labeled as tumors, while the remaining 59,135 were labeled as non-tumorous tissues. The evaluation of the image dataset revealed that the wide-ResNet model had the highest accuracy of 0.768, with an F1 score of 0.766. A preliminary evaluation on one patient indicated alignment between the ground truth set by neurosurgeons, the model’s predictions, and histopathological findings. Our deep learning algorithm has a positive tumor discrimination performance in intraoperative 4-K endoscopic images in patients with pituitary adenomas.

DOI： 10.1007/s10143-023-02196-w

Web of Science

Scopus

PubMed

Language：English   Publisher：Scientific Reports  

Efforts to optimize known materials and enhance their performance are ongoing, driven by the advancements resulting from the discovery of novel functional materials. Traditionally, the search for and optimization of functional materials has relied on the experience and intuition of specialized researchers. However, materials informatics (MI), which integrates materials data and machine learning, has frequently been used to realize systematic and efficient materials exploration without depending on manual tasks. Nonetheless, the discovery of new materials using MI remains challenging. In this study, we propose a method for the discovery of materials outside the scope of existing databases by combining MI with the experience and intuition of researchers. Specifically, we designed a two-dimensional map that plots known materials data based on their composition and structure, facilitating researchers’ intuitive search for new materials. The materials map was implemented using an autoencoder-based neural network. We focused on the conductivity of 708 lithium oxide materials and considered the correlation with migration energy (ME), an index of lithium-ion conductivity. The distribution of existing data reflected in the materials map can contribute to the development of new lithium-ion conductive materials by enhancing the experience and intuition of material researchers.

DOI： 10.1038/s41598-023-43921-1

Open Access

Web of Science

Scopus

PubMed

Language：English   Publisher：Journal of Biological Chemistry  

Cutting-edge technologies such as genome editing and synthetic biology allow us to produce novel foods and functional proteins. However, their toxicity and allergenicity must be accurately evaluated. It is known that specific amino acid sequences in proteins make some proteins allergic, but many of these sequences remain uncharacterized. In this study, we introduce a data-driven approach and a machine-learning method to find undiscovered allergen-specific patterns (ASPs) among amino acid sequences. The proposed method enables an exhaustive search for amino acid subsequences whose frequencies are statistically significantly higher in allergenic proteins. As a proof-of-concept, we created a database containing 21,154 proteins of which the presence or absence of allergic reactions are already known and applied the proposed method to the database. The detected ASPs in this proof-of-concept study were consistent with known biological findings, and the allergenicity prediction performance using the detected ASPs was higher than extant approaches, indicating this method may be useful in evaluating the utility of synthetic foods and proteins.

DOI： 10.1016/j.jbc.2023.104733

Open Access

Web of Science

Scopus

PubMed

Authorship：Last author,　Corresponding author   Language：English   Publishing type：Research paper (international conference proceedings)  

Language：English   Publishing type：Research paper (international conference proceedings)  

Language：English   Publisher：Medical Image Analysis  

In the present study, we propose a novel case-based similar image retrieval (SIR) method for hematoxylin and eosin (H&E) stained histopathological images of malignant lymphoma. When a whole slide image (WSI) is used as an input query, it is desirable to be able to retrieve similar cases by focusing on image patches in pathologically important regions such as tumor cells. To address this problem, we employ attention-based multiple instance learning, which enables us to focus on tumor-specific regions when the similarity between cases is computed. Moreover, we employ contrastive distance metric learning to incorporate immunohistochemical (IHC) staining patterns as useful supervised information for defining appropriate similarity between heterogeneous malignant lymphoma cases. In the experiment with 249 malignant lymphoma patients, we confirmed that the proposed method exhibited higher evaluation measures than the baseline case-based SIR methods. Furthermore, the subjective evaluation by pathologists revealed that our similarity measure using IHC staining patterns is appropriate for representing the similarity of H&E stained tissue images for malignant lymphoma.

DOI： 10.1016/j.media.2023.102752

Open Access

Web of Science

Scopus

PubMed

Language：English   Publishing type：Research paper (scientific journal)  

DOI： 10.1109/TPAMI.2022.3167993

Web of Science

Scopus

PubMed

Language：English   Publishing type：Research paper (scientific journal)  

DOI： 10.1007/s10463-022-00837-3

Web of Science

Scopus

Language：English   Publishing type：Research paper (international conference proceedings)  

Language：English   Publishing type：Research paper (international conference proceedings)  

Authorship：Last author,　Corresponding author   Language：English   Publishing type：Research paper (scientific journal)  

Language：English   Publishing type：Research paper (international conference proceedings)  

Language：English   Publishing type：Research paper (international conference proceedings)  

Publisher：Machine Learning  

Conformal prediction constructs a confidence set for an unobserved response of a feature vector based on previous identically distributed and exchangeable observations of responses and features. It has a coverage guarantee at any nominal level without additional assumptions on their distribution. Its computation deplorably requires a refitting procedure for all replacement candidates of the target response. In regression settings, this corresponds to an infinite number of model fits. Apart from relatively simple estimators that can be written as pieces of linear function of the response, efficiently computing such sets is difficult, and is still considered as an open problem. We exploit the fact that, often, conformal prediction sets are intervals whose boundaries can be efficiently approximated by classical root-finding algorithms. We investigate how this approach can overcome many limitations of formerly used strategies; we discuss its complexity and drawbacks.

DOI： 10.1007/s10994-022-06233-5

Scopus

Language：English   Publishing type：Research paper (scientific journal)  

DOI： 10.1007/s10463-022-00846-2

Web of Science

Scopus

Authorship：Last author,　Corresponding author   Language：English   Publishing type：Research paper (international conference proceedings)  

Authorship：Last author,　Corresponding author   Language：English   Publishing type：Research paper (scientific journal)  

DOI： 10.1162/neco_a_01550

PubMed

Language：English   Publishing type：Research paper (scientific journal)  

DOI： 10.1038/s41420-022-01232-w

Open Access

Authorship：Last author,　Corresponding author   Language：English   Publishing type：Research paper (scientific journal)  

Language：English   Publisher：Rsc Advances  

NASICON-type LiZr<inf>2</inf>(PO<inf>4</inf>)<inf>3</inf> (LZP) has attracted significant attention as a solid oxide electrolyte for all-solid-state Li-ion or Li-metal batteries owing to its high Li-ion conductivity, usability in all-solid-state batteries, and electrochemical stability against Li metal. In this study, we aim to improve the Li-ion conductivity of Li-rich NASICON-type LZPs doped with CaO and SiO<inf>2</inf>, i.e., Li<inf>1+x+2y</inf>Ca<inf>y</inf>Zr<inf>2−y</inf>Si<inf>x</inf>P<inf>3−x</inf>O<inf>12</inf>(0 ≤ x ≤ 0.3, 0 ≤ y ≤ 0.3) (LCZSP). Herein, a total of 49 compositions were synthesised, and their crystal structures, relative densities, and Li-ion conductivities were characterised experimentally. We confirmed the improvement in Li-ion conductivity by simultaneous replacement of Zr and P sites with Ca and Si ions, respectively. However, the intuition-derived determination of the composition exhibiting the highest Li-ion conductivity is technically difficult because the compositional dependence of the relative density and the crystalline phase of the sample is very complex. Bayesian optimisation (BO) was performed to efficiently discover the optimal composition that exhibited the highest Li-ion conductivity among the samples evaluated experimentally. We also optimised the composition of the LCZSP using multi-task Gaussian process regression after transferring prior knowledge of 47 compositions of Li<inf>1+x+2y</inf>Y<inf>x</inf>Ca<inf>y</inf>Zr<inf>2−x−y</inf>P<inf>3</inf>O<inf>12</inf> (0 ≤ x ≤ 0.376, 0 ≤ y ≤ 0.376) (LYCZP), i.e., BO with transfer learning. The present study successfully demonstrated that BO with transfer learning can search for optimal compositions two times as rapid as the conventional BO approach. This approach can be widely applicable for the optimisation of various functional materials as well as ionic conductors.

DOI： 10.1039/d2ra04539g

Open Access

Web of Science

Scopus

PubMed

Language：English   Publisher：Proceedings of the National Academy of Sciences of the United States of America  

Cyclones can cause mass mortality of seabirds, sometimes wrecking thousands of individuals. The few studies to track pelagic seabirds during cyclones show they tend to circumnavigate the strongest winds. We tracked adult shearwaters in the Sea of Japan over 11 y and found that the response to cyclones varied according to the wind speed and direction. In strong winds, birds that were sandwiched between the storm and mainland Japan flew away from land and toward the eye of the storm, flying within ≤30 km of the eye and tracking it for up to 8 h. This exposed shearwaters to some of the highest wind speeds near the eye wall (≤21 m s^–1) but enabled them to avoid strong onshore winds in the storm’s wake. Extreme winds may therefore become a threat when an inability to compensate for drift could lead to forced landings and collisions. Birds may need to know where land is in order to avoid it. This provides additional selective pressure for a map sense and could explain why juvenile shearwaters, which lack a map sense, instead navigating using a compass heading, are susceptible to being wrecked. We suggest that the ability to respond to storms is influenced by both flight and navigational capacities. This may become increasingly pertinent due to changes in extreme weather patterns.

DOI： 10.1073/pnas.2212925119

Web of Science

Scopus

PubMed

Language：English   Publishing type：Research paper (scientific journal)  

DOI： 10.1162/neco_a_01530

Open Access

Web of Science

Scopus

PubMed

Language：English   Publishing type：Research paper (scientific journal)  

DOI： 10.1002/pssb.202100525

Scopus

Language：English   Publishing type：Research paper (scientific journal)  

DOI： 10.1039/d2cc01526a

Open Access

Language：English   Publishing type：Research paper (scientific journal)  

DOI： 10.1162/neco_a_01530

Language：English   Publisher：Chemical Communications  

All-solid-state Li-ion batteries are of considerable interest as safer alternatives to Li-ion batteries containing flammable organic electrolytes. To date, however, achieving sufficient charging and discharging rates, in addition to capacity, at room temperature using these all-solid-state batteries has been challenging. To overcome these issues, material simulations and informatics investigations of a relatively new Na superionic conductor (NASICON)-type LiZr<inf>2</inf>(PO<inf>4</inf>)<inf>3</inf> (LZP) electrolyte were conducted to elucidate its characteristics and material functions. The following thermodynamic and/or kinetic properties of NASICON-type Li-ion conductive oxides were investigated with respect to the crystal structure mainly using material simulation and informatics approaches: (1) the electrochemical stabilities of LZP materials with respect to Li metal and (2) Li-ion conductivities in the bulk and at the grain boundaries. An efficient materials informatics search method was employed to optimise the material functions of the LZP electrolyte via Bayesian optimisation. This study should promote the application of LZP in all-solid-state batteries for use in technologies such as mobile devices and electric vehicles and enable more complex composition and process control.

DOI： 10.1039/d2cc01526a

Open Access

Web of Science

Scopus

PubMed

Language：English   Publishing type：Research paper (scientific journal)  

DOI： 10.1007/s11548-021-02549-0

Open Access

Web of Science

Scopus

PubMed

Authorship：Last author,　Corresponding author   Language：English   Publishing type：Research paper (international conference proceedings)  

Language：English   Publisher：Npj Parkinson S Disease  

To elucidate the relevance of gut dysbiosis in Parkinson’s disease (PD) in disease progression, we made random forest models to predict the progression of PD in two years by gut microbiota in 165 PD patients. The area under the receiver operating characteristic curves (AUROCs) of gut microbiota-based models for Hoehn & Yahr (HY) stages 1 and 2 were 0.799 and 0.705, respectively. Similarly, gut microbiota predicted the progression of Movement Disorder Society-Unified Parkinson’s Disease Rating Scale (MDS-UPDRS) III scores in an early stage of PD with AUROC = 0.728. Decreases of short-chain fatty acid-producing genera, Fusicatenibacter, Faecalibacterium, and Blautia, as well as an increase of mucin-degrading genus Akkermansia, predicted accelerated disease progression. The four genera remained unchanged in two years in PD, indicating that the taxonomic changes were not the consequences of disease progression. PD patients with marked gut dysbiosis may thus be destined to progress faster than those without gut dysbiosis.

DOI： 10.1038/s41531-022-00328-5

Open Access

Web of Science

Scopus

PubMed

Authorship：Last author,　Corresponding author   Language：English   Publishing type：Research paper (international conference proceedings)