Peer‐Timo Bremer · OSTI OAI (U.S. Department of Energy Office of Scientific and Technical Information) · 2023
A system for identifying objects in an image is provided. The system identifies segments of an image that may contain objects. For each segment, the system generates a segment score by inputting to a multi-scale neural network windows of multiple scales that include the segment that have been resampled to a fixed window size. A multi-scale neural network includes a feature extracting convolutional neural network (“feCNN”) for each scale and a classifier that inputs each feature of each feCNN. The segment score indicates whether the segment contains an object. The system generates a pixel score for pixels of the image. The pixel score for a pixel indicates that that pixel is within an object based on the segment scores of segments that contain that pixel. The system then identifies the object based on the pixel scores of neighboring pixels.
Cited by 0Open access
Jun Zhang, Yao-Kun Lei, Yi Yang et al. · ChemRxiv · 2020
Molecular simulations are widely applied in the study of chemical and bio-physical systems. However, theaccessible timescales of atomistic simulations are limited, and extracting equilibrium properties of systemscontaining rare events remains challenging. Two distinct strategies are usually adopted in this regard: eithersticking to the atomistic level and performing enhanced sampling, or trading details for speed by leveragingcoarse-grained models. Although both strategies are promising, either of them, if adopted individually,exhibits severe limitations. In this paper we propose a machine-learning approach to ally both strategies sothat simulations on different scales can benefit mutually from their cross-talks: Accurate coarse-grained (CG)models can be inferred from the fine-grained (FG) simulations through deep generative learning; In turn, FGsimulations can be boosted by the guidance of CG models via deep reinforcement learning. Our methoddefines a variational and adaptive training objective which allows end-to-end training of parametricmolecular models using deep neural networks. Through multiple experiments, we show that our method isefficient and flexible, and performs well on challenging chemical and bio-molecular systems.
Cited by 0Open access
Huaying Wang, Qiwen Li, Shuo Wang et al. · Applied Optics · 2024
Aimed at the problems of complex operation and zero-order influence in the reconstruction of holograms of multi-scale and large objects, a hologram reconstruction method based on deep learning was proposed. The UHC-Net model combined with the ConvNeXt module and hierarchical polymerization HANC module was adopted as the network structure to achieve accurate reconstruction of holograms. Using analog holograms and actual digital holograms as mixed data sets for reconstruction experiments, the generalization of the network can be extended. The results show that the proposed method can improve the reconstruction speed while maintaining the resolution and accuracy of reconstructed images. The results of this study can be applied to the reconstruction of large-size holograms, and the network generalization ability can be improved by using an analog data set, which provides a useful reference for the application of digital holography detection and deep learning in optical imaging.
Cited by 4
Kenneth Tran, Ashkan Panahi, Aniruddha Adiga et al. · 2019
We propose a deep learning architecture capable of performing up to 8× single image super-resolution. Our architecture incorporates an adversarial component from the super-resolution generative adversarial networks (SRGANs) and a multi-scale learning component from the multiple scale super-resolution network (MSSRNet), which only together can recover smaller structures inherent in satellite images. To further enhance our performance, we integrate progressive growing and training to our network. This, aided by feed forwarding connections in the network to move along and enrich information from previous inputs, produces super-resolved images at scaling factors of 2, 4, and 8. To ensure and enhance the stability of GANs, we employ Wasserstein GANs (WGANs) during training. Experimentally, we find that our architecture can recover small objects in satellite images during super-resolution whereas previous methods cannot.
Cited by 7
Tanveer Ahmad, Huanxin Chen · Energy · 2019
Cited by 87
Jun Zhang, Yao-Kun Lei, Yi Yang et al. · ChemRxiv · 2020
Molecular simulations are widely applied in the study of chemical and bio-physical systems. However, theaccessible timescales of atomistic simulations are limited, and extracting equilibrium properties of systemscontaining rare events remains challenging. Two distinct strategies are usually adopted in this regard: eithersticking to the atomistic level and performing enhanced sampling, or trading details for speed by leveragingcoarse-grained models. Although both strategies are promising, either of them, if adopted individually,exhibits severe limitations. In this paper we propose a machine-learning approach to ally both strategies sothat simulations on different scales can benefit mutually from their cross-talks: Accurate coarse-grained (CG)models can be inferred from the fine-grained (FG) simulations through deep generative learning; In turn, FGsimulations can be boosted by the guidance of CG models via deep reinforcement learning. Our methoddefines a variational and adaptive training objective which allows end-to-end training of parametricmolecular models using deep neural networks. Through multiple experiments, we show that our method isefficient and flexible, and performs well on challenging chemical and bio-molecular systems.
Cited by 0Open access
Victor Boussange, Bert Wuyts, Philipp Brun et al. · ArXiv.org · 2025
Biodiversity assessments depend critically on the spatial scale at which species richness is measured. How species richness accumulates with sampling area is influenced by natural and anthropogenic processes whose effects vary across spatial scales. These accumulation dynamics, described by the species-area relationship (SAR), are challenging to assess because most biodiversity surveys cover sampling areas far smaller than the scales at which these processes operate. Here, we combine sampling theory with deep learning to estimate species richness at arbitrary spatial scales across geographic space from existing ecological surveys. We apply our model, named MuScaRi, to ~350k vegetation surveys across Europe. Validated against independent regional plant inventories, MuScaRi reduces root mean squared error of vascular plant richness estimates by 61% relative to conventional estimators, yields substantially less biased predictions, and produces multi-scale richness maps alongside spatially explicit estimates of the species accumulation rate, a key indicator for biodiversity conservation. By encompassing the full spectrum of ecologically relevant spatial scales within a single unified framework, MuScaRi provides an essential tool for robust biodiversity assessments and forecasts under global change.
Cited by 0Open access
Liu, Xiaoyi · Scholar Commons (University of South Carolina) · 2024
Synthetic biology advances and combines the expertise of engineers and biologists, bridging the gap between engineering and natural life. Synthetic biology has been generally categorized into two broad branches by developing new biological components, networks, and systems to reprogram organisms. The first branch involves using synthetic molecules to mimic natural biological functions. The second branch focuses on assembling natural biological components in novel ways, aiming to produce systems with unique, practical functions. Thus, the de novo engineering of biological modules and synthetic pathways is used in related practical bioengineering applications, such as drug-targeting strategies and microbial product manufacturing. Therefore, synthetic biology represents a new paradigm in scientific exploration and innovation, with widely used implications for our understanding and optimization of biological systems. Over the past decades, there has been a significant increase in the amount of available whole-genome sequencing data and experimental data due to the emergence of new automation technologies, such as high-content imaging, high-throughput screening, and sequencing. Given the growth of these data sets, researchers are unable to summarize these data simply from experience and memory. Thus, stable and efficient computational methods are required to integrate them to predict or reveal new phenomena or insights that have never been discovered. However, incomplete knowledge of metabolic processes impairs the accuracy of biological systems, hindering advancements in systems biology and metabolic engineering. Additionally, some fundamental challenges still remain. Firstly, problems in systems biology are often cross-scale and multi-modal, yet existing computational methods for problem definition and model design are often single-scale and single-modal. Secondly, biological systems are multi-scale, unbalanced, and noisy, making structuring and benchmarking this complicated data very difficult. Thirdly, most natural or valuable products' complete biosynthetic pathways are unknown. Thus, computer-aided biosynthesis planning holds significant value. To address the above challenges, we introduce multi-scale deep learning-based representation learning methodologies to understand and optimize the downstream tasks in systems biology, such as metabolic pathway inference, missing reaction prediction in GEMs, and retrosynthesis prediction. Specifically, our first study introduces a novel Multi-View Multi-Label learning framework for Metabolic Pathway Inference (MVML-MPI), which outperforms State-Of-The-Art (SOTA) methods by accurately representing the complex relationships between compounds and pathways. In the second study, to address the limitation of incomplete metabolic knowledge in GEMs, we proposed a novel framework named hypergraph Convolution network and attention mechanism integrated Explorer for GAPS prediction of metabolism termed CLOSEgaps. It is a comprehensive deep learning-driven tool that represents the hyper-topological information of GEMs and effectively fills gaps through hyperlink prediction, thereby enhancing the accuracy of phenotypic predictions. In the third study, we proposed a novel end-to-end framework for one-step retrosynthesis that combines the power of a graph encoder, which integrates learnable structural information, with the capability to sequentially translate drugs, thereby efficiently capturing chemically plausible information (RetroCaptioner). This research presents an advancement in systems biology by introducing a suite of multi-scale deep learning methodologies. These methodologies tackle key challenges such as MVML-MPI enhancing our understanding of complex metabolic pathways, CLOSEgaps innovatively filling gaps in metabolic models, and RetroCaptioner facilitating the process of retrosynthesis. Taken together, they form a comprehensive and integrated approach, and our proposed methods significantly advance the capabilities of synthetic biology.
Cited by 0
Jiao Wang, Bin Wu, Zhenwen Ren et al. · IEEE Access · 2022
Deep subspace clustering methods have achieved impressive clustering performance compared with other clustering algorithms. However, most existing methods suffer from the following problems: 1) they only consider the global features but neglect the local features in subspace self-expressiveness learning; 2) they neglect the discriminative information of each self-expressiveness coefficient matrix; 3) they ignore the useful long-range dependencies and positional information in feature representation learning. To solve these problems, in this paper, we propose a novel multi-scale deep subspace clustering with discriminative learning (MDSCDL) to obtain a high-quality self-expressiveness coefficient matrix. Specifically, MDSCDL bridges multiple fully-connection layers between encoder and decoder to learn multi-scale self-expressiveness coefficient matrices from global and local features, representing the more comprehensive relationship among data. By modeling the interdependencies of the multi-scale self-expressiveness coefficient matrices, MDSCDL adaptively assigns discriminative weights for each matrix and fuses them with convolution operation. Moreover, to increase representation power, MDSCDL introduces the coordinate attention mechanism to extract the long-range dependencies and positional features for subspace self-expressiveness learning. Extensive experiments on the face and object datasets have shown the superiority of MDSCDL compared with several state-of-the-art methods.
Cited by 11Open access
Jun Zhang, Yao-Kun Lei, Yi Yang et al. · ChemRxiv · 2019
Molecular simulations are widely applied in the study of chemical and bio-physical systems. However, the<br>accessible timescales of atomistic simulations are limited, and extracting equilibrium properties of systems<br>containing rare events remains challenging. Two distinct strategies are usually adopted in this regard: either<br>sticking to the atomistic level and performing enhanced sampling, or trading details for speed by leveraging<br>coarse-grained models. Although both strategies are promising, either of them, if adopted individually,<br>exhibits severe limitations. In this paper we propose a machine-learning approach to ally both strategies so<br>that simulations on different scales can benefit mutually from their cross-talks: Accurate coarse-grained (CG)<br>models can be inferred from the fine-grained (FG) simulations through deep generative learning; In turn, FG<br>simulations can be boosted by the guidance of CG models via deep reinforcement learning. Our method<br>defines a variational and adaptive training objective which allows end-to-end training of parametric<br>molecular models using deep neural networks. Through multiple experiments, we show that our method is<br>efficient and flexible, and performs well on challenging chemical and bio-molecular systems. <br>
Cited by 3Open access
Thorsten Behrens, Karsten Schmidt, R.A. MacMillan et al. · Scientific Reports · 2018
Cited by 147Open access
Quentin Potié, Guillaume Touya, Chaimaa Beladraoui et al. · Abstracts of the ICA · 2022
International audience
Cited by 3Open access
Xingyi Zhang, Ran Cheng, Ye Tian et al. · 2024
This chapter presents two state-of-the-art Evolutionary algorithms (EAs) for two typical multi-objective optimization problems for deep learning: a gradient-guided multi-objective EA for training deep neural networks (DNNs) (termed GEMONN) and an action command-based surrogate-assisted multi-objective EA for searching neural architectures of DNNs (termed ACEncoding). Moreover, the network sparsity is also optimized together with the training loss, which can reduce the network complexity and alleviate overfitting. Experimental results on single-layer NNs, deep-layer NNs, recurrent NNs, and CNNs demonstrate the effectiveness of the proposed GEMONN. In short, this work not only introduces a novel approach for training deep NNs but also enhances the performance of EAs in solving large-scale optimization problems. For the proposed ACEncoding, a novel encoding scheme and a performance evaluator are presented for neural architecture search.
Cited by 0
Sruthi T P, Helen Mary M C · 2025
Cardiovascular diseases remain the leading cause of mortality worldwide, the development of accurate and automated diagnostic systems to address the growing burden on healthcare systems. Electrocardiogram signals, widely used for cardiac monitoring, are often prone to noise and require robust preprocessing and analysis techniques for reliable arrhythmia detection. This research introduces a novel framework that leverages Discrete Wavelet Transform for effective ECG signal denoising and Continuous Wavelet Transform for seamless ID-to-2D signal conversion. These enhanced signal representations are fed into a hybrid deep learning architecture by combining EfficientNet and ResNet to improve arrhythmia classification accuracy. By integrating these advanced methods, the proposed system aims to achieve high diagnostic accuracy while maintaining computational efficiency. The approach is evaluated on the widely recognized MIT-BIH Arrhythmia Database, covering multiple arrhythmia types to ensure robust performance. The hybrid EfficientNet-ResNet model achieves 97% test accuracy in arrhythmia classification. This study underscores the potential of hybrid deep learning frame works in revolutionizing cardiac care by providing accurate, automated, and efficient arrhythmia detection systems, ultimately contributing to improved patient outcomes and reduced workload for healthcare professionals.
Cited by 0
Guangle Yao, Tao Leí, Jiandan Zhong et al. · Applied Intelligence · 2018
Cited by 28
Zhenzhong Kuang, Zongmin Li, Tianyi Zhao et al. · 2017
To achieve more effective solution for large-scale image classification (i.e., classifying millions of images into thousands or even tens of thousands of object classes or categories), a deep multi-task learning algorithm is developed by seamlessly integrating deep CNNs with multi-task learning over the concept ontology, where the concept ontology is used to organize large numbers of object classes or categories hierarchically and determine the inter-related learning tasks automatically. Our deep multi-task learning algorithm can integrate the deep CNNs to learn more discriminative high-level features for image representation, and it can also leverage multi-task learning and inter-level relationship constraint to train more discriminative tree classifier over the concept ontology and control the inter-level error propagation effectively. In our deep multi-task learning algorithm, we can use back propagation to simultaneously refine both the relevant node classifiers (at different levels of the concept ontology) and the deep CNNs according to a joint objective function. The experimental results have demonstrated that our deep multi-task learning algorithm can achieve very competitive results on both the accuracy and the cost of feature extraction for large-scale image classification.
Cited by 25
Xuelin Qian, Yanwei Fu, Yu–Gang Jiang et al. · 2017
Person Re-identification (re-id) aims to match people across non-overlapping camera views in a public space. It is a challenging problem because many people captured in surveillance videos wear similar clothes. Consequently, the differences in their appearance are often subtle and only detectable at the right location and scales. Existing re-id models, particularly the recently proposed deep learning based ones match people at a single scale. In contrast, in this paper, a novel multi-scale deep learning model is proposed. Our model is able to learn deep discriminative feature representations at different scales and automatically determine the most suitable scales for matching. The importance of different spatial locations for extracting discriminative features is also learned explicitly. Experiments are carried out to demonstrate that the proposed model outperforms the state-of-the art on a number of benchmarks.
Cited by 297
Jun Zhu, Zeyang Zhang, Zeyang Zhang et al. · bioRxiv (Cold Spring Harbor Laboratory) · 2024
Abstract Cells are the fundamental structural and functional units of life. Studying the definition and composition of different cell types can help us understand the complex mechanisms underlying biological diversity and functionality. The increasing volume of extensive single-cell omics data makes it possible to provide detailed characterisations of cell types. Recently, there has been a rise in deep learning-based approaches that generate cell type labels solely through mapping query data to reference data. However, these approaches lack multi-scale descriptions and interpretations of identified cell types. Here, we propose Cell Decoder, a biological prior knowledge informed model to achieve multi-scale representation of cells. We implemented automated machine learning and post-hoc analysis techniques to decode cell identity. We have shown that Cell Decoder compares favourably to existing methods, offering multi-view interpretability for decoding cell identity and data integration. Furthermore, we have showcased its applicability in uncovering novel cell types and states in both human bone and mouse embryonic contexts, thereby revealing the multi-scale heterogeneity inherent in cell identities.
Cited by 3Open access
Fu, Y, Qian, X, Jiang, Y-G et al. · OPUS - Open Publications of UTS Scholars (University of Technology Sydney) · 2017
Person Re-identification (re-id) aims to match people across non-overlapping camera views in a public space. It is a challenging problem because many people captured in surveillance videos wear similar clothes. Consequently, the differences in their appearance are often subtle and only detectable at the right location and scales. Existing re-id models, particularly the recently proposed deep learning based ones match people at a single scale. In contrast, in this paper, a novel multi-scale deep learning model is proposed. Our model is able to learn deep discriminative feature representations at different scales and automatically determine the most suitable scales for matching. The importance of different spatial locations for extracting discriminative features is also learned explicitly. Experiments are carried out to demonstrate that the proposed model outperforms the state-of-the art on a number of benchmarks.
Cited by 3
Natalia Neverova, Christian Wolf, Graham W. Taylor et al. · Lecture notes in computer science · 2015
Cited by 223