Receive a weekly summary and discussion of the top papers of the week by leading researchers in the field.

General General

Immune Signature Against Plasmodium falciparum Antigens Predicts Clinical Immunity in Distinct Malaria Endemic Communities.

In Molecular & cellular proteomics : MCP

A large body of evidence supports the role of antibodies directed against the Plasmodium spp. parasite in the development of naturally acquired immunity to malaria, however an antigen signature capable of predicting protective immunity against Plasmodium remains to be identified. Key challenges for the identification of a predictive immune signature include the high dimensionality of data produced by high-throughput technologies and the limitation of standard statistical tests in accounting for synergetic interactions between immune responses to multiple targets. In this study, using samples collected from young children in Ghana at multiple time points during a longitudinal study, we adapted a predictive modeling framework which combines feature selection and machine learning techniques to identify an antigen signature of clinical immunity to malaria. Our results show that an individual's immune status can be accurately predicted by measuring antibody responses to a small defined set of 15 target antigens. We further demonstrate that the identified immune signature is highly versatile and capable of providing precise and accurate estimates of clinical protection from malaria in an independent geographic community. Our findings pave the way for the development of a robust point-of-care test to identify individuals at high risk of disease and which could be applied to monitor the impact of vaccinations and other interventions. This approach could be also translated to biomarker discovery for other infectious diseases.

Proietti Carla, Krause Lutz, Trieu Angela, Dodoo Daniel, Gyan Ben, Koram Kwadwo A, Rogers William O, Richie Thomas L, Crompton Peter D, Felgner Philip L, Doolan Denise L


Biomarker: diagnostic, Plasmodium falciparum, antigen signature, clinical data, feature selection, immunology, machine learning, malaria, modeling, protein microarray

General General

Neural Networks for Estimating Speculative Attacks Models.

In Entropy (Basel, Switzerland)

Currency crises have been analyzed and modeled over the last few decades. These currency crises develop mainly due to a balance of payments crisis, and in many cases, these crises lead to speculative attacks against the price of the currency. Despite the popularity of these models, they are currently shown as models with low estimation precision. In the present study, estimates are made with first- and second-generation speculative attack models using neural network methods. The results conclude that the Quantum-Inspired Neural Network and Deep Neural Decision Trees methodologies are shown to be the most accurate, with results around 90% accuracy. These results exceed the estimates made with Ordinary Least Squares, the usual estimation method for speculative attack models. In addition, the time required for the estimation is less for neural network methods than for Ordinary Least Squares. These results can be of great importance for public and financial institutions when anticipating speculative pressures on currencies that are in price crisis in the markets.

Alaminos David, Aguilar-Vijande Fernando, Sánchez-Serrano José Ramón


Quantum-Inspired Neural Network, currency crisis, deep learning, neural networks, speculative attacks

Pathology Pathology

Deep learning powers cancer diagnosis in digital pathology.

In Computerized medical imaging and graphics : the official journal of the Computerized Medical Imaging Society

Technological innovation has accelerated the pathological diagnostic process for cancer, especially in digitizing histopathology slides and incorporating deep learning-based approaches to mine the subvisual morphometric phenotypes for improving pathology diagnosis. In this perspective paper, we provide an overview on major deep learning approaches for digital pathology and discuss challenges and opportunities of such approaches to aid cancer diagnosis in digital pathology. In particular, the emerging graph neural network may further improve the performance and interpretability of deep learning in digital pathology.

He Yunjie, Zhao Hong, Wong Stephen T C


AI, Digital pathology, cancer diagnosis, deep learning, graph neural networks, microscopy image

General General

Changes in human-nature relations during pandemic outbreaks: a big data analysis.

In The Science of the total environment

Pandemic outbreaks can cause diverse impacts on society by altering human-nature relations. This study analyzed these relational changes during the severe acute respiratory syndrome (SARS), Swine flu, Middle East respiratory syndrome (MERS), and Ebola outbreaks by applying machine learning and big data analyses of global news articles. The results showed that social-ecological systems play vital roles in analyzing indirect pandemic impacts. Herein, major pandemic impacts, including reduced use of cultural ecosystem services, can be analyzed by big data analyses at the global scale. All the identified pandemic impacts herein were linked to provisioning and cultural ecosystem services, implying that these ecosystem services might be more recognized or valued more by the public than regulating and supporting ecosystem services. Further, the pandemic impacts were presented with human-centric views, indicating a challenge to adapting nature-based solutions to mitigate the risk of future pandemic emergences. These findings will advance the current knowledge of diverse pandemic impacts and human-nature relations.

Jaung Wanggi


Coronaviruses, Ecosystem services, Media analysis, Natural language processing, Social-ecological systems

General General

PM2.5 concentration forecasting at surface monitoring sites using GRU neural network based on empirical mode decomposition.

In The Science of the total environment

The main component of haze is the particulate matter (PM) 2.5. How to explore the laws of PM2.5 concentration changes is the main content of air quality prediction. Combining the characteristics of temporality and non-linearity in PM2.5 concentration series, more and more deep learning methods are currently applied to PM2.5 predictions, but most of them ignore the non-stationarity of time series, which leads to a lower accuracy of model prediction. To address this issue, an integration method of gated recurrent unit neural network based on empirical mode decomposition (EMD-GRU) for predicting PM2.5 concentration was proposed in this paper. This method uses empirical mode decomposition (EMD) to decompose the PM2.5 concentration sequence first and then fed the multiple stationary sub-sequences obtained after the decomposition and the meteorological features into the constructed GRU neural network successively for training and predicting. Finally, the sub-sequences of the prediction output are added to obtain the prediction results of PM2.5 concentration. The forecast result of the case in this paper show that the EMD-GRU model reduces the RMSE by 44%, MAE by 40.82%, and SMAPE by 11.63% compared to the single GRU model.

Huang Guoyan, Li Xinyi, Zhang Bing, Ren Jiadong


Deep learning, Empirical mode decomposition, Gated recurrent unit neural network, PM2.5 concentration prediction, Time series

General General

Stacked DeBERT: All attention in incomplete data for text classification.

In Neural networks : the official journal of the International Neural Network Society

In this paper, we propose Stacked DeBERT, short for StackedDenoising Bidirectional Encoder Representations from Transformers. This novel model improves robustness in incomplete data, when compared to existing systems, by designing a novel encoding scheme in BERT, a powerful language representation model solely based on attention mechanisms. Incomplete data in natural language processing refer to text with missing or incorrect words, and its presence can hinder the performance of current models that were not implemented to withstand such noises, but must still perform well even under duress. This is due to the fact that current approaches are built for and trained with clean and complete data, and thus are not able to extract features that can adequately represent incomplete data. Our proposed approach consists of obtaining intermediate input representations by applying an embedding layer to the input tokens followed by vanilla transformers. These intermediate features are given as input to novel denoising transformers which are responsible for obtaining richer input representations. The proposed approach takes advantage of stacks of multilayer perceptrons for the reconstruction of missing words' embeddings by extracting more abstract and meaningful hidden feature vectors, and bidirectional transformers for improved embedding representation. We consider two datasets for training and evaluation: the Chatbot Natural Language Understanding Evaluation Corpus and Kaggle's Twitter Sentiment Corpus. Our model shows improved F1-scores and better robustness in informal/incorrect texts present in tweets and in texts with Speech-to-Text error in the sentiment and intent classification tasks.1.

Cunha Sergio Gwenaelle, Lee Minho


BERT, Denoising, Incomplete data, Incomplete text classification, Speech-to-Text error, Transformers