In the context of lung ultrasound, the detection of B-lines, which are indicative of interstitial lung disease and pulmonary edema, plays a pivotal role in clinical diagnosis. Current methods still rely on visual inspection by experts. Vision-based automatic B-line detection methods have been developed, but their performance has yet to improve in terms of both accuracy and computational speed. This paper presents a novel approach to posing B-line detection as an inverse problem via deep unfolding of the Alternating Direction Method of Multipliers (ADMM). It tackles the challenges of data labelling and model training in lung ultrasound image analysis by harnessing the capabilities of deep neural networks and model-based methods. Our objective is to substantially enhance diagnostic accuracy while ensuring efficient real-time capabilities. The results show that the proposed method runs more than 90 times faster than the traditional model-based method and achieves an F1 score that is 10.6% higher.
Studies have proved that the number of B-lines in lung ultrasound images has a strong statistical link to the amount of extravascular lung water, which is significant for hemodialysis treatment. Manual inspection of B-lines requires experts and is time-consuming, whilst modelling automation methods is currently problematic because of a lack of ground truth. Therefore, in this paper, we propose a novel semi-supervised learning method for the B-line detection task based on contrastive learning. Through multi-level unsupervised learning on unlabelled lung ultrasound images, the features of the artefacts are learnt. In the downstream task, we introduce a fine-tuning process on a small number of labelled images using the EIoU-based loss function. Apart from reducing the data labelling workload, the proposed method shows a superior performance to model-based algorithm with the recall of 91.43%, the accuracy of 84.21% and the F1 score of 91.43%.
Sponsored search ads appear next to search results when people look for products and services on search engines. In recent years, they have become one of the most lucrative channels for marketing. As the fundamental basis of search ads, relevance modeling has attracted increasing attention due to the significant research challenges and tremendous practical value. Most existing approaches solely rely on the semantic information in the input query-ad pair, while the pure semantic information in the short ads data is not sufficient to fully identify user's search intents. Our motivation lies in incorporating the tremendous amount of unsupervised user behavior data from the historical search logs as the complementary graph to facilitate relevance modeling. In this paper, we extensively investigate how to naturally fuse the semantic textual information with the user behavior graph, and further propose three novel AdsGNN models to aggregate topological neighborhood from the perspectives of nodes, edges and tokens. Furthermore, two critical but rarely investigated problems, domain-specific pre-training and long-tail ads matching, are studied thoroughly. Empirically, we evaluate the AdsGNN models over the large industry dataset, and the experimental results of online/offline tests consistently demonstrate the superiority of our proposal.
In the field of biomedical imaging, ultrasonography has become increasingly widespread, and an important auxiliary diagnostic tool with unique advantages, such as being non-ionising and often portable. This article reviews the state-of-the-art in medical ultrasound image computing and in particular its application in the examination of the lungs. First, we review the current developments in medical ultrasound technology. We then focus on the characteristics of lung ultrasonography and on its ability to diagnose a variety of diseases through the identification of various artefacts. We review medical ultrasound image processing methods by splitting them into two categories: (1) traditional model-based methods, and (2) data driven methods. For the former, we consider inverse problem based methods by focusing in particular on ultrasound image despeckling, deconvolution, and line artefacts detection. Among the data-driven approaches, we discuss various works based on deep/machine learning, which include various effective network architectures implementing supervised, weakly supervised and unsupervised learning.
Text encoders based on C-DSSM or transformers have demonstrated strong performance in many Natural Language Processing (NLP) tasks. Low latency variants of these models have also been developed in recent years in order to apply them in the field of sponsored search which has strict computational constraints. However these models are not the panacea to solve all the Natural Language Understanding (NLU) challenges as the pure semantic information in the data is not sufficient to fully identify the user intents. We propose the TextGNN model that naturally extends the strong twin tower structured encoders with the complementary graph information from user historical behaviors, which serves as a natural guide to help us better understand the intents and hence generate better language representations. The model inherits all the benefits of twin tower models such as C-DSSM and TwinBERT so that it can still be used in the low latency environment while achieving a significant performance gain than the strong encoder-only counterpart baseline models in both offline evaluations and online production system. In offline experiments, the model achieves a 0.14% overall increase in ROC-AUC with a 1% increased accuracy for long-tail low-frequency Ads, and in the online A/B testing, the model shows a 2.03% increase in Revenue Per Mille with a 2.32% decrease in Ad defect rate.
Deep learning using Convolutional Neural Networks (CNNs) has been shown to significantly out-performed many conventional vision algorithms. Despite efforts to increase the CNN efficiency both algorithmically and with specialized hardware, deep learning remains difficult to deploy in resource-constrained environments. In this paper, we propose an end-to-end framework to explore optically compute the CNNs in free-space, much like a computational camera. Compared to existing free-space optics-based approaches which are limited to processing single-channel (i.e., grayscale) inputs, we propose the first general approach, based on nanoscale meta-surface optics, that can process RGB data directly from the natural scenes. Our system achieves up to an order of magnitude energy saving, simplifies the sensor design, all the while sacrificing little network accuracy.