China's export growth continued to rise in July, sending trade surplus to a record high, according to government data. China's exports grew 18% to $333 billion compared to the same period last year, and were up from 17.9% in June, according to data from China's customs. Imports, however, remained soft, growing 2.3% in July compared to a year ago. That was lower that economists' estimates of 4%, and suggests weak domestic demand amid lockdowns across the country as China attempts to stem the outbreak of COVID-19. China's total trade surplus reached an all-time high of $101.3 billion in July, breaking the record set in June.
Then there is the ever-present pandemic that has been threatening this planet for years. It got me thinking: can technology be used to combat all these horrible diseases and improve patient outcomes. Is artificial intelligence going to play a role in this? We've achieved another milestone in Artificial Intelligence adoption: $6.9 Billion of market value and counting. The intelligent healthcare market will reach 67.4 Billion by 2027. The future of AI in healthcare is bright but not peaceful.
Martin Lukac, associate professor from Nazarbayev University (NU) School of Engineering and Digital Sciences, and Alessandra Clementi, assistant professor from NU School of Medicine, explore how cough recordings and AI will be used to diagnose patients in the CoughAnalyzer app. Cough is used as an early symptom for a variety of pulmonary diseases and can be used to track the progress of a respiratory disease or infection. Since the COVID-19 pandemic, cough analysis has been popularised for early symptoms indicators that were able to distinguish between COVID-19 and other types of coughs. However, in many cases, the data is regional with high variation and high-quality differences, so the usage of cough analysis still remains in development. More generally, remote cough analysis has the advantage of allowing diagnosis at a distance and providing early evidence of a potentially more serious problem.
"For regression, a voting ensemble involves making a prediction that is the average of multiple other regression models. In classification, a hard voting ensemble involves summing the votes for crisp class labels from other models and predicting the class with the most votes. A soft voting ensemble involves summing the predicted probabilities for class labels and predicting the class label with the largest sum probability." The principle is that you are leveraging the strengths of each model's predictive ability -- rather than relying on a single model's prediction. It's important to note that ensemble voting is only effective if the models in the ensemble are meaningfully different in how they learn on the data. If I design and build 5 models that are exactly the same (same architecture and trained on the same data) and ask the 5 models to vote, they probably wouldn't produce results that are meaningfully different from having the data pass through just one of those models. It's actually arguably worse since it would be 5x computationally more expensive for a similar or exact same result. If you were to read any of the papers linked above, you'll notice that Ensemble Voting algorithms are actually starting to get used in conjunction with transfer learning. That is researchers are beginning to apply transfer learning methods to pneumonia problems and then passing the results through various ensemble voting algorithms to produce some of the best results to-date.
Classification labels specifying the presence and absence of abnormalities are necessary to train computer vision models on radiology images. However, obtaining these classification labels by hand is time-consuming and limits the size of the final dataset as well as the number of abnormalities that can be considered. In this post we'll overview an easily customizable technique, SARLE, for extracting structured abnormality and location labels automatically from the free-text reports that accompany each radiology image in a hospital database. Every time a patient is imaged, a radiologist interprets the image and writes a report summarizing the normal and abnormal findings. An excerpt from a chest x-ray report might read, "There is a nodule in the right lung. The left lung is clear. There is cardiomegaly without pericardial effusion."
Artificial intelligence, or AI, is ubiquitous and integrated into almost any field or application. Progress made over the last few decades has been astounding, with achievements--such as DeepMind's AlphaGo defeating the world's foremost Go player in 2016 and the application of LinearFold to predict the secondary structure of the SARS-CoV-2 RNA sequence in just 26 seconds--demonstrating the ever-growing capabilities of these systems. There is a caveat though; running AI requires large amounts of energy and data, and computer hardware can't keep up. Integrated circuit chips are reaching capacity even as structures on the chips and circuit components become smaller. There is a limit to how far we can physically take this.
The predictive capability of artificial intelligence (AI) machine learning is accelerating discoveries in life science. A new study shows how AI and genomics can predict future mutations of the SARS-CoV-2 virus that causes the COVID-19 disease. "The severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) pandemic has been characterized by waves of transmission initiated by new variants replacing older ones," wrote the Broad Institute of MIT and Harvard research team, with their co-authors from the University of Massachusetts Medical School and other affiliations. "Given this pattern of emergence, there is an obvious need for the early detection of novel variants to prevent excess deaths." The research team developed a hierarchical Bayesian regression AI model called PyR0 that can provide scalable analytics of the complete set of public datasets of SARS-CoV-2 genomes. The Bayesian model predicts emerging viral lineages.
AGI solutions are being continuously investigated, though the current most promising mainstream technology, neural networks, while contributing to some extraordinary results, are still running short of achieving them. This criticism is not new, and, most recently Gary Marcus, in "Deep Learning: A Critical Appraisal", arXiv:1801.00631v1, has outlined many issues with current deep learning architectures, in particular their inability to'understand' the information they manipulate and their ability to mostly work in a'stable' world. As Marcus states in his article: 'The logic of deep learning is such that it is likely to work best in highly stable worlds, like the board game Go, which has unvarying rules, and less well in systems such as politics and economics that are constantly changing. To the extent that deep learning is applied in tasks such as stock prediction, there is a good chance that it will eventually face the fate of Google Flu Trends, which initially did a great job of predicting epidemological [sic] data on search trends, only to complete [sic] miss things like the peak of the 2013 flu season (Lazer, Kennedy, King & Vespignani, 2014)'. Even one of the so called'fathers' of Deep Learning architectures, Geoffrey Hinton, has recently voiced his concerns that deep learning needs to start over.
A neck patch that monitors respiratory sounds may help manage asthma and chronic obstructive pulmonary disease (COPD) by detecting symptom flareups in real time, without compromising patient privacy. Asthma and COPD are two of the most common chronic respiratory diseases. In Europe, the combined prevalence is about 10 percent of the general population. In Canada, an estimated 3.8 million people experience asthma and two million people experience COPD. The chronic nature of asthma and COPD requires continuous disease monitoring and management.