scholarly journals ScaleNet - Improve CNNs through Recursively Rescaling Objects

2020 ◽  
Vol 34 (07) ◽  
pp. 11426-11433
Author(s):  
Xingyi Li ◽  
Zhongang Qi ◽  
Xiaoli Fern ◽  
Fuxin Li
Keyword(s):  
Deep Learning ◽  
Large Scale ◽  
Testing Time ◽  
Learning Models ◽  
Training Set ◽  
Scale Invariant ◽  
Learning Framework ◽  
Deep Networks ◽  
Training Sets

Deep networks are often not scale-invariant hence their performance can vary wildly if recognizable objects are at an unseen scale occurring only at testing time. In this paper, we propose ScaleNet, which recursively predicts object scale in a deep learning framework. With an explicit objective to predict the scale of objects in images, ScaleNet enables pretrained deep learning models to identify objects in the scales that are not present in their training sets. By recursively calling ScaleNet, one can generalize to very large scale changes unseen in the training set. To demonstrate the robustness of our proposed framework, we conduct experiments with pretrained as well as fine-tuned classification and detection frameworks on MNIST, CIFAR-10, and MS COCO datasets and results reveal that our proposed framework significantly boosts the performances of deep networks.

scholarly journals EyeHealer: A large-scale anterior eye segment dataset with eye structure and lesion annotations

2021 ◽  
Author(s):  
Wenjia Cai ◽  
Jie Xu ◽  
Ke Wang ◽  
Xiaohong Liu ◽  
Wenqin Xu ◽  
...  
Keyword(s):  
Deep Learning ◽  
Large Scale ◽  
Image Annotation ◽  
Clinical Care ◽  
Significant Proportion ◽  
Anterior Segment ◽  
Future Research ◽  
Learning Models ◽  
Anterior Eye Segment ◽  
Segmentation Models

Abstract Anterior segment eye diseases account for a significant proportion of presentations to eye clinics worldwide, including diseases associated with corneal pathologies, anterior chamber abnormalities (e.g. blood or inflammation) and lens diseases. The construction of an automatic tool for the segmentation of anterior segment eye lesions will greatly improve the efficiency of clinical care. With research on artificial intelligence progressing in recent years, deep learning models have shown their superiority in image classification and segmentation. The training and evaluation of deep learning models should be based on a large amount of data annotated with expertise, however, such data are relatively scarce in the domain of medicine. Herein, the authors developed a new medical image annotation system, called EyeHealer. It is a large-scale anterior eye segment dataset with both eye structures and lesions annotated at the pixel level. Comprehensive experiments were conducted to verify its performance in disease classification and eye lesion segmentation. The results showed that semantic segmentation models outperformed medical segmentation models. This paper describes the establishment of the system for automated classification and segmentation tasks. The dataset will be made publicly available to encourage future research in this area.

scholarly journals A Physics-Infused Deep Learning Model for the Prediction of Refractive Indices and Its Use for the Large-Scale Screening of Organic Compound Space

2019 ◽  
Author(s):  
Mojtaba Haghighatlari ◽  
Gaurav Vishwakarma ◽  
Mohammad Atif Faiz Afzal ◽  
Johannes Hachmann
Keyword(s):  
Machine Learning ◽  
Deep Learning ◽  
Large Scale ◽  
Organic Molecules ◽  
Learning Model ◽  
Training Data ◽  
Refractive Indices ◽  
Learning Models ◽  
Deep Learning Model ◽  
Machine Learning Models

<div><div><div><p>We present a multitask, physics-infused deep learning model to accurately and efficiently predict refractive indices (RIs) of organic molecules, and we apply it to a library of 1.5 million compounds. We show that it outperforms earlier machine learning models by a significant margin, and that incorporating known physics into data-derived models provides valuable guardrails. Using a transfer learning approach, we augment the model to reproduce results consistent with higher-level computational chemistry training data, but with a considerably reduced number of corresponding calculations. Prediction errors of machine learning models are typically smallest for commonly observed target property values, consistent with the distribution of the training data. However, since our goal is to identify candidates with unusually large RI values, we propose a strategy to boost the performance of our model in the remoter areas of the RI distribution: We bias the model with respect to the under-represented classes of molecules that have values in the high-RI regime. By adopting a metric popular in web search engines, we evaluate our effectiveness in ranking top candidates. We confirm that the models developed in this study can reliably predict the RIs of the top 1,000 compounds, and are thus able to capture their ranking. We believe that this is the first study to develop a data-derived model that ensures the reliability of RI predictions by model augmentation in the extrapolation region on such a large scale. These results underscore the tremendous potential of machine learning in facilitating molecular (hyper)screening approaches on a massive scale and in accelerating the discovery of new compounds and materials, such as organic molecules with high-RI for applications in opto-electronics.</p></div></div></div>

scholarly journals Mol-BERT: An Effective Molecular Representation with BERT for Molecular Property Prediction

2021 ◽  
Vol 2021 ◽  
pp. 1-7
Author(s):  
Juncai Li ◽  
Xiaofei Jiang
Keyword(s):  
Deep Learning ◽  
Language Processing ◽  
Large Scale ◽  
Molecular Data ◽  
Molecular Property ◽  
Property Prediction ◽  
Learning Framework ◽  
Learning Techniques ◽  
Potential Benefits ◽  
Current Sequence

Molecular property prediction is an essential task in drug discovery. Most computational approaches with deep learning techniques either focus on designing novel molecular representation or combining with some advanced models together. However, researchers pay fewer attention to the potential benefits in massive unlabeled molecular data (e.g., ZINC). This task becomes increasingly challenging owing to the limitation of the scale of labeled data. Motivated by the recent advancements of pretrained models in natural language processing, the drug molecule can be naturally viewed as language to some extent. In this paper, we investigate how to develop the pretrained model BERT to extract useful molecular substructure information for molecular property prediction. We present a novel end-to-end deep learning framework, named Mol-BERT, that combines an effective molecular representation with pretrained BERT model tailored for molecular property prediction. Specifically, a large-scale prediction BERT model is pretrained to generate the embedding of molecular substructures, by using four million unlabeled drug SMILES (i.e., ZINC 15 and ChEMBL 27). Then, the pretrained BERT model can be fine-tuned on various molecular property prediction tasks. To examine the performance of our proposed Mol-BERT, we conduct several experiments on 4 widely used molecular datasets. In comparison to the traditional and state-of-the-art baselines, the results illustrate that our proposed Mol-BERT can outperform the current sequence-based methods and achieve at least 2% improvement on ROC-AUC score on Tox21, SIDER, and ClinTox dataset.

Regularization strategies for deep-learning-based salt model building

2019 ◽  
Vol 7 (4) ◽  
pp. T911-T922
Author(s):  
Satyakee Sen ◽  
Sribharath Kainkaryam ◽  
Cen Ong ◽  
Arvind Sharma
Keyword(s):  
Deep Learning ◽  
Large Scale ◽  
Model Building ◽  
Ground Truth ◽  
Real Data ◽  
Test Time ◽  
Generalization Error ◽  
Training Set ◽  
Production Scale ◽  
Ensemble Strategy

Salt model building has long been considered a severe bottleneck for large-scale 3D seismic imaging projects. It is one of the most time-consuming, labor-intensive, and difficult-to-automate processes in the entire depth imaging workflow requiring significant intervention by domain experts to manually interpret the salt bodies on noisy, low-frequency, and low-resolution seismic images at each iteration of the salt model building process. The difficulty and need for automating this task is well-recognized by the imaging community and has propelled the use of deep-learning-based convolutional neural network (CNN) architectures to carry out this task. However, significant challenges remain for reliable production-scale deployment of CNN-based methods for salt model building. This is mainly due to the poor generalization capabilities of these networks. When used on new surveys, never seen by the CNN models during the training stage, the interpretation accuracy of these models drops significantly. To remediate this key problem, we have introduced a U-shaped encoder-decoder type CNN architecture trained using a specialized regularization strategy aimed at reducing the generalization error of the network. Our regularization scheme perturbs the ground truth labels in the training set. Two different perturbations are discussed: one that randomly changes the labels of the training set, flipping salt labels to sediments and vice versa and the second that smooths the labels. We have determined that such perturbations act as a strong regularizer preventing the network from making highly confident predictions on the training set and thus reducing overfitting. An ensemble strategy is also used for test time augmentation that is shown to further improve the accuracy. The robustness of our CNN models, in terms of reduced generalization error and improved interpretation accuracy is demonstrated with real data examples from the Gulf of Mexico.

scholarly journals Progressive System: A Deep-Learning Framework for Real-Time Data in Industrial Production

Processes ◽  
2020 ◽  
Vol 8 (6) ◽  
pp. 649
Author(s):  
Yifeng Liu ◽  
Wei Zhang ◽  
Wenhao Du
Keyword(s):  
Deep Learning ◽  
Real Time ◽  
Large Scale ◽  
Quality Data ◽  
Time Data ◽  
High Quality ◽  
Real Time System ◽  
High Quality Data ◽  
Learning Framework ◽  
Data Accumulation

Deep learning based on a large number of high-quality data plays an important role in many industries. However, deep learning is hard to directly embed in the real-time system, because the data accumulation of the system depends on real-time acquisitions. However, the analysis tasks of such systems need to be carried out in real time, which makes it impossible to complete the analysis tasks by accumulating data for a long time. In order to solve the problems of high-quality data accumulation, high timeliness of the data analysis, and difficulty in embedding deep-learning algorithms directly in real-time systems, this paper proposes a new progressive deep-learning framework and conducts experiments on image recognition. The experimental results show that the proposed framework is effective and performs well and can reach a conclusion similar to the deep-learning framework based on large-scale data.

scholarly journals Deploying Machine and Deep Learning Models for Efficient Data-Augmented Detection of COVID-19 Infections

Viruses ◽  
2020 ◽  
Vol 12 (7) ◽  
pp. 769 ◽  
Author(s):  
Ahmed Sedik ◽  
Abdullah M Iliyasu ◽  
Basma Abd El-Rahiem ◽  
Mohammed E. Abdel Samea ◽  
Asmaa Abdel-Raheem ◽  
...  
Keyword(s):  
Deep Learning ◽  
Data Augmentation ◽  
Short Term Memory ◽  
Machine Learning Techniques ◽  
Detection Accuracy ◽  
Testing Time ◽  
Learning Models ◽  
Accurate Identification ◽  
Infected People ◽  
Corona Virus

This generation faces existential threats because of the global assault of the novel Corona virus 2019 (i.e., COVID-19). With more than thirteen million infected and nearly 600000 fatalities in 188 countries/regions, COVID-19 is the worst calamity since the World War II. These misfortunes are traced to various reasons, including late detection of latent or asymptomatic carriers, migration, and inadequate isolation of infected people. This makes detection, containment, and mitigation global priorities to contain exposure via quarantine, lockdowns, work/stay at home, and social distancing that are focused on “flattening the curve”. While medical and healthcare givers are at the frontline in the battle against COVID-19, it is a crusade for all of humanity. Meanwhile, machine and deep learning models have been revolutionary across numerous domains and applications whose potency have been exploited to birth numerous state-of-the-art technologies utilised in disease detection, diagnoses, and treatment. Despite these potentials, machine and, particularly, deep learning models are data sensitive, because their effectiveness depends on availability and reliability of data. The unavailability of such data hinders efforts of engineers and computer scientists to fully contribute to the ongoing assault against COVID-19. Faced with a calamity on one side and absence of reliable data on the other, this study presents two data-augmentation models to enhance learnability of the Convolutional Neural Network (CNN) and the Convolutional Long Short-Term Memory (ConvLSTM)-based deep learning models (DADLMs) and, by doing so, boost the accuracy of COVID-19 detection. Experimental results reveal improvement in terms of accuracy of detection, logarithmic loss, and testing time relative to DLMs devoid of such data augmentation. Furthermore, average increases of 4% to 11% in COVID-19 detection accuracy are reported in favour of the proposed data-augmented deep learning models relative to the machine learning techniques. Therefore, the proposed algorithm is effective in performing a rapid and consistent Corona virus diagnosis that is primarily aimed at assisting clinicians in making accurate identification of the virus.

scholarly journals Validating the validation: reanalyzing a large-scale comparison of deep learning and machine learning models for bioactivity prediction

2020 ◽  
Vol 34 (7) ◽  
pp. 717-730 ◽  
Author(s):  
Matthew C. Robinson ◽  
Robert C. Glen ◽  
Alpha A. Lee
Keyword(s):  
Machine Learning ◽  
Deep Learning ◽  
Numerical Experiments ◽  
Large Scale ◽  
Operating Characteristic ◽  
Characteristic Curve ◽  
Learning Models ◽  
Bioactivity Prediction ◽  
Operating Characteristic Curve ◽  
Machine Learning Models

Abstract Machine learning methods may have the potential to significantly accelerate drug discovery. However, the increasing rate of new methodological approaches being published in the literature raises the fundamental question of how models should be benchmarked and validated. We reanalyze the data generated by a recently published large-scale comparison of machine learning models for bioactivity prediction and arrive at a somewhat different conclusion. We show that the performance of support vector machines is competitive with that of deep learning methods. Additionally, using a series of numerical experiments, we question the relevance of area under the receiver operating characteristic curve as a metric in virtual screening. We further suggest that area under the precision–recall curve should be used in conjunction with the receiver operating characteristic curve. Our numerical experiments also highlight challenges in estimating the uncertainty in model performance via scaffold-split nested cross validation.

scholarly journals Visualizing Deep Networks by Optimizing with Integrated Gradients

2020 ◽  
Vol 34 (07) ◽  
pp. 11890-11898
Author(s):  
Zhongang Qi ◽  
Saeed Khorram ◽  
Li Fuxin
Keyword(s):  
Deep Learning ◽  
State Of The Art ◽  
User Needs ◽  
Learning Models ◽  
Local Optima ◽  
Popular Approach ◽  
Deep Network ◽  
Deep Networks ◽  
Benchmark Datasets ◽  
Descent Directions

Understanding and interpreting the decisions made by deep learning models is valuable in many domains. In computer vision, computing heatmaps from a deep network is a popular approach for visualizing and understanding deep networks. However, heatmaps that do not correlate with the network may mislead human, hence the performance of heatmaps in providing a faithful explanation to the underlying deep network is crucial. In this paper, we propose I-GOS, which optimizes for a heatmap so that the classification scores on the masked image would maximally decrease. The main novelty of the approach is to compute descent directions based on the integrated gradients instead of the normal gradient, which avoids local optima and speeds up convergence. Compared with previous approaches, our method can flexibly compute heatmaps at any resolution for different user needs. Extensive experiments on several benchmark datasets show that the heatmaps produced by our approach are more correlated with the decision of the underlying deep network, in comparison with other state-of-the-art approaches.

scholarly journals A Novel Object-Based Deep Learning Framework for Semantic Segmentation of Very High-Resolution Remote Sensing Data: Comparison with Convolutional and Fully Convolutional Networks

2019 ◽  
Vol 11 (6) ◽  
pp. 684 ◽  
Author(s):  
Maria Papadomanolaki ◽  
Maria Vakalopoulou ◽  
Konstantinos Karantzalos
Keyword(s):  
Deep Learning ◽  
State Of The Art ◽  
Semantic Segmentation ◽  
Novel Object ◽  
Convolutional Networks ◽  
Learning Framework ◽  
Fully Convolutional Networks ◽  
Object Based ◽  
Deep Networks ◽  
Very High

Deep learning architectures have received much attention in recent years demonstrating state-of-the-art performance in several segmentation, classification and other computer vision tasks. Most of these deep networks are based on either convolutional or fully convolutional architectures. In this paper, we propose a novel object-based deep-learning framework for semantic segmentation in very high-resolution satellite data. In particular, we exploit object-based priors integrated into a fully convolutional neural network by incorporating an anisotropic diffusion data preprocessing step and an additional loss term during the training process. Under this constrained framework, the goal is to enforce pixels that belong to the same object to be classified at the same semantic category. We compared thoroughly the novel object-based framework with the currently dominating convolutional and fully convolutional deep networks. In particular, numerous experiments were conducted on the publicly available ISPRS WGII/4 benchmark datasets, namely Vaihingen and Potsdam, for validation and inter-comparison based on a variety of metrics. Quantitatively, experimental results indicate that, overall, the proposed object-based framework slightly outperformed the current state-of-the-art fully convolutional networks by more than 1% in terms of overall accuracy, while intersection over union results are improved for all semantic categories. Qualitatively, man-made classes with more strict geometry such as buildings were the ones that benefit most from our method, especially along object boundaries, highlighting the great potential of the developed approach.

Sign in / Sign up

Export Citation Format

Share Document