13197 research outputs found
Sort by
Topological image analysis and manipulation
No full textvideo/mp4Many digital image applications require extracting structural information from digital images and preserving/repairing those structures when images are treated. Here, we focus on topological information such as connected components, cavities and tunnels (in 3D). Digital topology and mathematical morphology have been involved in various studies on this subject, while persistent homology is recently used in various fields in order to study the shape and structure of data. In this talk, I will briefly present the three approaches with some comparative study in the context of image analysis and manipulation, and show their utility with some application examples.講演日: 2023年6月8日 4限講演場所: 情報科学棟 エーアイ大講義室(L1)vide
エヴァンゲリオン ノ ジレイ ニ ミル キャラクター ライセンス ウンヨウ ノ ゲンジョウ
No full textvideo/mp4全世界に人気の広がる日本のアニメーションは、そのライセンス活用においても多様さと新奇性で作品の付加価値を向上させてきた。多くの話題を喚起してきた『エヴァンゲリオン』シリーズの多岐にわたる企業コラボレーションや商品化によるプロモーション事例を紹介し、知的財産の活用の可能性を展望する。講演日: 2023年6月15日 4限講演場所: 情報科学棟 エーアイ大講義室(L1)vide
ダイキボ データ ヲ カツヨウ シテ ブンシ ヲ セッケイ スル
No full textvideo/mp4低分子創薬においては大量のデータを有効に利用するための「方法や手法」が長年研究されてきましたが、近年のAI技術の急速な発展により「方法や手法」が洗練化されてきました。この講座では、コンピュータを利用した低分子創薬において、分子構造をどのようにデータに基づいて設計するか、その適用限界や可能性について考えます。講演場所: ミレニアムホール講演日: 2023年10月21日本学物質創成科学領域vide
Generative Replay Inspired by Hippocampal Memory Indexing for Continual Language Learning
No full textContinual learning aims to accumulate knowledge to solve new tasks without catastrophic forgetting for previously learned tasks. Research on continual learning has led to the development of generative replay, which prevents catastrophic forgetting by generating pseudo-samples for previous tasks and learning them together with new tasks. Inspired by the biological brain, we propose the hippocampal memory indexing to enhance the generative replay by controlling sample generation using compressed features of previous training samples. It enables the generation of a specific training sample from previous tasks, thus improving the balance and quality of generated replay samples. Experimental results indicate that our method effectively controls the sample generation and consistently outperforms the performance of current generative replay methods.conference pape
Spatial attention and quantization-based contrastive learning framework for mmWave massive MIMO beam training
No full textDeep learning (DL)-based beam training schemes have been exploited to improve spectral efficiency with fast optimal beam selection for millimeter-wave (mmWave) massive multiple-input multiple-output (MIMO) systems. To achieve high prediction accuracy, these DL models rely on training with a tremendous amount of labeled environmental measurements, such as mmWave channel state information (CSI). However, demanding a large volume of ground truth labels for beam training is inefficient and infeasible due to the high labeling cost and the requirement for expertise in practical mmWave massive MIMO systems. Meanwhile, a complex environment incurs critical performance degradation in the continuous output of beam training. In this paper, we propose a novel contrastive learning framework, named self-enhanced quantized phase-based transformer network (SE-QPTNet), for reliable beam training with only a small fraction of the labeled CSI dataset. We first develop a quantized phase-based transformer network (QPTNet) with a hierarchical structure to explore the essential features from frequency and spatial views and quantize the environmental components with a latent beam codebook to achieve robust representation. Next, we design the SE-QPTNet including self-enhanced pre-training and supervised beam training. SE-QPTNet pre-trains by the contrastive information of the target user and others with the unlabeled CSI, and then, it is utilized as the initialization to fine-tune with a reduced volume of labeled CSI. Finally, the experimental results show that the proposed framework improves beam prediction accuracy and data rates with 5% labeled data compared to existing solutions. Our proposed framework further enhances flexibility and breaks the limitation of the quantity of label information for practical beam training.journal articl
Arabidopsis BBX14 negatively regulates nitrogen starvation- and dark-induced leaf senescence
No full textSenescence is a highly regulated process driven by developmental age and environmental factors. Although leaf senescence is accelerated by nitrogen (N) deficiency, the underlying physiological and molecular mechanisms are largely unknown. Here, we reveal that BBX14, a previously uncharacterized BBX-type transcription factor in Arabidopsis, is crucial for N starvation-induced leaf senescence. We find that inhibiting BBX14 by artificial miRNA (amiRNA) accelerates senescence during N starvation and in darkness, while BBX14 overexpression (BBX14-OX) delays it, identifying BBX14 as a negative regulator of N starvation- and dark-induced senescence. During N starvation, nitrate and amino acids like glutamic acid, glutamine, aspartic acid, and asparagine were highly retained in BBX14-OX leaves compared to the wild type. Transcriptome analysis showed a large number of senescence-associated genes (SAGs) to be differentially expressed between BBX14-OX and wild-type plants, including ETHYLENE INSENSITIVE3 (EIN3) which regulates N signaling and leaf senescence. Chromatin immunoprecipitation (ChIP) showed that BBX14 directly regulates EIN3 transcription. Furthermore, we revealed the upstream transcriptional cascade of BBX14. By yeast one-hybrid screen and ChIP, we found that MYB44, a stress-responsive MYB transcription factor, directly binds to the promoter of BBX14 and activates its expression. In addition, Phytochrome Interacting Factor 4 (PIF4) binds to the promoter of BBX14 to repress BBX14 transcription. Thus, BBX14 functions as a negative regulator of N starvation-induced senescence through EIN3 and is directly regulated by PIF4 and MYB44.journal articl
Cross-Species Metabolomic Analyses in the Brassicaceae Reveals Common Responses to Ultraviolet-B Exposure
No full textExposure to UV-B radiation, an intrinsic component of solar light, is detrimental to all living organisms as chromophore units of DNA, RNA and proteins readily absorb high-energy photons. Indirect damage to the same molecules and lipids is mediated by elevated reactive oxygen species (ROS) levels, a side effect of exposure to UV-B stress. To protect themselves from UV-B radiation, plants produce phytochemical sunscreens, among which flavonoids have shown to be particularly effective. The core aglycone of flavonoid molecules is subjected to chemical decoration, such as glycosylation and acylation, further improving sunscreen properties. In particular, acylation, which adds a phenolic ring to flavonoid molecules, enhances the spectral absorption of UV-A and UV-B rays, providing to this class of compounds exceptional shielding power. In this study, we comprehensively analyzed the responses to UV-B radiation in four Brassicaceae species, including Arabidopsis thaliana, Brassica napus, Brassica oleracea, and Brassica rapa. Our study revealed a complete reprogramming of the central metabolic pathway in response to UV-B radiation characterized by increased production of functional precursors of specialized metabolites with UV-B shielding properties, indicating a targeted effort of plant metabolism to provide increased protection. The analysis of specialized metabolites and transcripts revealed the activation of the phenylpropanoid$2013acetate pathway, leading to the production of specific classes of flavonoids and a cross-species increase in phenylacylated-flavonoid glucosides with synapoyl glycoside decorations. Interestingly, our analysis also revealed that acyltransferase genes of the class of serine carboxypeptidase-like (SCPLs) proteins are costitutively expressed, but downregulated in response to UV-B radiation, possibly independently of the ELONGATED HYPOCOTYL 5 (HY5) signaling pathway.journal articl
Chemometrics Approach Based on Wavelet Transforms for the Estimation of Monomer Concentrations from FTIR Spectra
No full textFourier-transform infrared (FTIR) spectroscopy can detect the presence of functional groups and molecules directly from a mixed solution of organic molecules. Although it is quite useful to monitor chemical reactions, quantitative analysis of FTIR spectra becomes difficult when various peaks of different widths overlap. To overcome this difficulty, we propose a chemometrics approach to accurately predict the concentration of components in chemical reactions, yet interpretable by humans. The proposed method first decomposes a spectrum into peaks with various widths by the wavelet transform. Subsequently, a sparse linear regression model is built using the wavelet coefficients. Models by the method are interpretable using the regression coefficients shown on Gaussian distributions with various widths. The interpretation is expected to reveal the relation of broad regions in spectra to the model prediction. In this study, we conducted the prediction of monomer concentration in copolymerization reactions of five monomers against methyl methacrylate by various chemometric approaches including conventional methods. A rigorous validation scheme revealed that the proposed method overall showed better predictive ability than various linear and non-linear regression methods. The visualization results were consistent with the interpretation obtained by another chemometric approach and qualitative evaluation. The proposed method is found to be useful for calculating the concentrations of monomers in copolymerization reactions and for the interpretation of spectra.journal articl
LATTE: Lattice ATTentive Encoding for Character-based Word Segmentation
No full textA character sequence comprises at least one or more segmentation alternatives. This can be considered segmentation ambiguity and may weaken segmentation performance in word segmentation. Proper handling of such ambiguity lessens ambiguous decisions on word boundaries. Previous works have achieved remarkable segmentation performance and alleviated the ambiguity problem by incorporating the lattice, owing to its ability to capture segmentation alternatives, along with graph-based and pre-trained models. However, multiple granularity information, including character and word, in a lattice that encodes with such models may not be attentively exploited. To strengthen multi-granularity representations in a lattice, we propose the Lattice ATTentive Encoding (LATTE) method for character-based word segmentation. Our model employs the lattice structure to handle segmentation alternatives and utilizes graph neural networks along with an attention mechanism to attentively extract multi-granularity representation from the lattice for complementing character representations. Our experimental results demonstrated improvements in segmentation performance on the BCCWJ, CTB6, and BEST2010 datasets in three languages, particularly Japanese, Chinese, and Thai.journal articl
Scalable Empirical Dynamic Modeling With Parallel Computing and Approximate k-NN Search
No full textEmpirical Dynamic Modeling (EDM) is a mathematical framework for modeling and predicting non-linear time series data. Although EDM is increasingly adopted in various research fields, its application to large-scale data has been limited due to its high computational cost. This article presents kEDM, a high-performance implementation of EDM for analyzing large-scale time series datasets. kEDM adopts the Kokkos performance-portable programming model to efficiently run on both CPU and GPU while sharing a single code base. We also conduct hardware-specific optimization of performance-critical kernels. kEDM achieved up to 6.58× speedup in pairwise causal inference of real-world biology datasets compared to an existing EDM implementation. Furthermore, we integrate multiple approximate k-NN search algorithms into EDM to enable the analysis of extremely large datasets that were intractable with conventional EDM based on exhaustive k-NN search. EDM-based time series forecast enhanced with approximate k-NN search demonstrated up to 790× speedup compared to conventional Simplex projection with less than 1% increase in MAPE.journal articl