Improving deep ensemble vehicle classification by using selected adversarial samples

Implementation and evaluation of an additional GPT-4-based reviewer in PRISMA-based medical systematic literature reviews

BACKGROUND

PRISMA-based literature reviews require meticulous scrutiny of extensive textual data by multiple reviewers, which is associated with considerable human effort.

OBJECTIVE

To evaluate feasibility and reliability of using GPT-4 API as a complementary reviewer in systematic literature reviews based on the PRISMA framework.

METHODOLOGY

A systematic literature review on the role of natural language processing and Large Language Models (LLMs) in automatic patient-trial matching was conducted using human reviewers and an AI-based reviewer (GPT-4 API). A RAG methodology with LangChain integration was used to process full-text articles. Agreement levels between two human reviewers and GPT-4 API for abstract screening and between a single reviewer and GPT-4 API for full-text parameter extraction were evaluated.

RESULTS

An almost perfect GPT-human reviewer agreement in the abstract screening process (Cohen's kappa > 0.9) and a lower agreement in the full-text parameter extraction were observed.

CONCLUSION

As GPT-4 has performed on a par with human reviewers in abstract screening, we conclude that GPT-4 has an exciting potential of being used as a main screening tool for systematic literature reviews, replacing at least one of the human reviewers.

Ensemble of diverse deep neural networks with pseudo-labels for repayment prediction in social lending

In peer-to-peer (P2P) social lending, it is important to predict the repayment of borrowers. P2P lending data are generated in real-time, but most of them are pending to decide the repayment because the deadline is not yet expired. Adding the unexpired data with appropriate labels into the training set could improve the performance of a prediction model, but the pseudo-labels cannot be certainly precise. In this paper, we propose an ensemble classifier composed of diverse convolutional neural networks (CNNs) of GoogLeNet, ResNet and DenseNet for the repayment prediction in social lending with the pseudo-labels approximated by an uncertainty handling scheme. The additional data labeled by Dempster-Shafer fusion of two semi-supervised learning methods boost up training of various models of CNNs, which are combined by weighted voting. A diversity measure is applied to constructing a pool of different models of CNNs that extract the effective features in the social lending data with labeling noise and predict the borrower's loan status. The experiment with the real dataset of 855,502 cases from Lending Club confirms that the diverse ensemble combined with weighted voting achieves the highest performance and outperforms conventional methods.

Vehicle Type Classification using Hybrid Features and a Deep Neural Network

Currently, considerable research has been done in vehicle type classification, especially due to the success of deep learning in many image classification problems. In this research, a system incorporating hybrid features is proposed to improve the performance of vehicle type classification. The feature vectors are extracted from the pre-processed images using Gabor features, a histogram of oriented gradients and a local optimal oriented pattern. The hybrid set of features contains complementary information that could help discriminate between the classes better, further, an ant colony optimizer is utilized to reduce the dimension of the extracted feature vectors. Finally, a deep neural network is used to classify the types of vehicles in the images. The proposed approach was tested on the MIO vision traffic camera dataset and another more challenging real-world dataset consisting of videos of multiple lanes of a toll plaza. The proposed model showed an improvement in accuracy ranging from 0.28% to 8.68% in the MIO TCD dataset when compared to well-known neural network architectures.

A Dynamic Convolutional Neural Network Based Shared-Bike Demand Forecasting Model

Bike-sharing systems are becoming popular and generate a large volume of trajectory data. In a bike-sharing system, users can borrow and return bikes at different stations. In particular, a bike-sharing system will be affected by weather, the time period, and other dynamic factors, which challenges the scheduling of shared bikes. In this article, a new shared-bike demand forecasting model based on dynamic convolutional neural networks, called SDF , is proposed to predict the demand of shared bikes. SDF chooses the most relevant weather features from real weather data by using the Pearson correlation coefficient and transforms them into a two-dimensional dynamic feature matrix, taking into account the states of stations from historical data. The feature information in the matrix is extracted, learned, and trained with a newly proposed dynamic convolutional neural network to predict the demand of shared bikes in a dynamical and intelligent fashion. The phase of parameter update is optimized from three aspects: the loss function, optimization algorithm, and learning rate. Then, an accurate shared-bike demand forecasting model is designed based on the basic idea of minimizing the loss value. By comparing with classical machine learning models, the weight sharing strategy employed by SDF reduces the complexity of the network. It allows a high prediction accuracy to be achieved within a relatively short period of time. Extensive experiments are conducted on real-world bike-sharing datasets to evaluate SDF. The results show that SDF significantly outperforms classical machine learning models in prediction accuracy and efficiency.

Melanoma Recognition by Fusing Convolutional Blocks and Dynamic Routing between Capsules

Simple Summary

The early treatment of skin cancer can effectively reduce mortality rates. Recently, automatic melanoma diagnosis from skin images has gained attention, which was mainly encouraged by the well-known challenge developed by the International Skin Imaging Collaboration project. The majority of contestant submitted Convolutional Neural Network based solutions. However, this type of model presents disadvantages. As a consequence, Dynamic Routing between Capsules has been proposed to overcome such limitations. The aim of our proposal was to assess the advantages of combining both architectures. An extensive experimental study showed the proposal significantly outperformed state-of-the-art models, achieving 166% higher predictive performance compared to ResNet in non-dermoscopic images. In addition, the pixels activated during prediction were shown, which allows to assess the rationale to give such a conclusion. Finally, more research should be conducted in order to demonstrate the potential of this neural network architecture in other areas.

Abstract

Skin cancer is one of the most common types of cancers in the world, with melanoma being the most lethal form. Automatic melanoma diagnosis from skin images has recently gained attention within the machine learning community, due to the complexity involved. In the past few years, convolutional neural network models have been commonly used to approach this issue. This type of model, however, presents disadvantages that sometimes hamper its application in real-world situations, e.g., the construction of transformation-invariant models and their inability to consider spatial hierarchies between entities within an image. Recently, Dynamic Routing between Capsules architecture (CapsNet) has been proposed to overcome such limitations. This work is aimed at proposing a new architecture which combines convolutional blocks with a customized CapsNet architecture, allowing for the extraction of richer abstract features. This architecture uses high-quality 299×299×3 skin lesion images, and a hyper-tuning of the main parameters is performed in order to ensure effective learning under limited training data. An extensive experimental study on eleven image datasets was conducted where the proposal significantly outperformed several state-of-the-art models. Finally, predictions made by the model were validated through the application of two modern model-agnostic interpretation tools.

Selective oversampling approach for strongly imbalanced data

Challenges posed by imbalanced data are encountered in many real-world applications. One of the possible approaches to improve the classifier performance on imbalanced data is oversampling. In this paper, we propose the new selective oversampling approach (SOA) that first isolates the most representative samples from minority classes by using an outlier detection technique and then utilizes these samples for synthetic oversampling. We show that the proposed approach improves the performance of two state-of-the-art oversampling methods, namely, the synthetic minority oversampling technique and adaptive synthetic sampling. The prediction performance is evaluated on four synthetic datasets and four real-world datasets, and the proposed SOA methods always achieved the same or better performance than other considered existing oversampling methods.

Accelerating the experimental responses on cell behaviors: a long-term prediction of cell trajectories using Social Generative Adversarial Network

The incremented uptake provided by time-lapse microscopy in Organ-on-a-Chip (OoC) devices allowed increased attention to the dynamics of the co-cultured systems. However, the amount of information stored in long-time experiments may constitute a serious bottleneck of the experimental pipeline. Forward long-term prediction of cell trajectories may reduce the spatial–temporal burden of video sequences storage. Cell trajectory prediction becomes crucial especially to increase the trustworthiness in software tools designed to conduct a massive analysis of cell behavior under chemical stimuli. To address this task, we transpose here the exploitation of the presence of “social forces” from the human to the cellular level for motion prediction at microscale by adapting the potential of Social Generative Adversarial Network predictors to cell motility. To demonstrate the effectiveness of the approach, we consider here two case studies: one related to PC-3 prostate cancer cells cultured in 2D Petri dishes under control and treated conditions and one related to an OoC experiment of tumor-immune interaction in fibrosarcoma cells. The goodness of the proposed strategy has been verified by successfully comparing the distributions of common descriptors (kinematic descriptors and mean interaction time for the two scenarios respectively) from the trajectories obtained by video analysis and the predicted counterparts.

Deep Learning in the Biomedical Applications: Recent and Future Status

Deep neural networks represent, nowadays, the most effective machine learning technology in biomedical domain. In this domain, the different areas of interest concern the Omics (study of the genome—genomics—and proteins—transcriptomics, proteomics, and metabolomics), bioimaging (study of biological cell and tissue), medical imaging (study of the human organs by creating visual representations), BBMI (study of the brain and body machine interface) and public and medical health management (PmHM). This paper reviews the major deep learning concepts pertinent to such biomedical applications. Concise overviews are provided for the Omics and the BBMI. We end our analysis with a critical discussion, interpretation and relevant open challenges.