Breast cancer was the most commonly diagnosed cancer among women worldwide in 2020. Recently, several deep learning-based classification approaches have been proposed to screen breast cancer in mammograms. However, most of these approaches require additional detection or segmentation annotations. Meanwhile, some other image-level label-based methods often pay insufficient attention to lesion areas, which are critical for diagnosis. This study designs a novel deep-learning method for automatically diagnosing breast cancer in mammography, which focuses on the local lesion areas and only utilizes image-level classification labels. In this study, we propose to select discriminative feature descriptors from feature maps instead of identifying lesion areas using precise annotations. And we design a novel adaptive convolutional feature descriptor selection (AFDS) structure based on the distribution of the deep activation map. Specifically, we adopt the triangle threshold strategy to calculate a specific threshold for guiding the activation map to determine which feature descriptors (local areas) are discriminative. Ablation experiments and visualization analysis indicate that the AFDS structure makes the model easier to learn the difference between malignant and benign/normal lesions. Furthermore, since the AFDS structure can be regarded as a highly efficient pooling structure, it can be easily plugged into most existing convolutional neural networks with negligible effort and time consumption. Experimental results on two publicly available INbreast and CBIS-DDSM datasets indicate that the proposed method performs satisfactorily compared with state-of-the-art methods.

Mammography is one of the most commonly used methods for breast cancer screening. Most mammogram classification or segmentation models are trained with additional manual segmentation annotation data, which is difficult to obtain. Meanwhile, these models rarely take into account the fact that lesion features may vary at different scales. Moreover, most models have difficulty in quantifying the effect of single lesion regions on prediction results. To solve the above problems, we propose deep multiscale multi-instance networks with regional scoring. This model classifies mammograms automatically by image labels without using any other annotation information during the training procedure. The proposed multiscale structures include large scale, medium scale, and small scale, and they can extract features of mammograms from global to local. Because there may be more than one breast lesion region in a mammogram, the proposed multi-instance structure divides the image into multiple regions and selects few regions of interest to represent the whole image together. The regional scoring mechanism quantifies the contribution of each region to the prediction results, and thus, makes the results interpretable. In addition, the proposed method can roughly localize the lesion area by the score of each region. The experimental results on a publicly available dataset and a private dataset have shown that the proposed method achieved the highest AUC on both datasets compared with the state-of-the-art models.

Data augmentation is an effective technique to enrich the training data’s diversity and reduce the risk of overfitting. However, different datasets have distinct preferences on various augmentation techniques. Recently, automated data augmentation (auto-augmentation), which could engineer augmentation policy automatically, drew a growing interest. Previous auto-augmentation methods usually utilize a Density Matching strategy, which highly depends on a large data scale to ensure a precise policy evaluation. When facing small-scale datasets, it usually achieves an inferior performance. To address the problem, an improved method named Augmented Density Matching is proposed by augmenting the train data with policies uniformly sampled from a prior distribution, making the policy evaluation more precise. Moreover, we propose a unified search framework for data augmentation and neural architecture (USAA) by formulating the search processes with one formulation. As a result, both optimal augmentation policy and neural architecture could be obtained within one round of search process. Extensive experiments have been conducted on a bag of medical datasets with general small scales, and the results show that the proposed Augmented Density Matching and USAA can outperform the state-of-the-art auto-augmentation and AutoML approaches, respectively.

Mammography is one of the most widely used and effective ways to screen early breast cancer. Convolutional neural network-based methods have obtained promising results for automatic mammography diagnosis. However, most of those approaches ignore the relationship between global and local characteristics of mammograms and lose sight of the relationship between different views of a patient. This study designs a novel region label assignment strategy, which takes advantage of all regions in each mammogram from a patient by assigning different labels to different regions and calculating loss for each region separately. This approach enables the classifier to distinguish variable and tiny lesions in complex global conditions better. And only one case-level classification label is needed for diagnosing one patient (case). Moreover, the categories of mammogram datasets are always imbalanced. To address this problem, this study designs an AUC (area under the receiver operating characteristic curve)-based optimization method on mini-batch strategy. Experimental results on a constructed dataset and two publicly available datasets demonstrate that the proposed method performs satisfactorily compared with state-of-the-art mammogram classifiers. Visualization results show the proposed method can find out mammograms containing malignant lesions and illustrate the rough location of lesions.

Through the use of sarcastic sentences on social media, people can express their strong emotions. Therefore, the detection of sarcasm in social media has received more and more attention over the past years. Classifying a sentence as sarcastic or nonsarcastic heavily relies on the contextual information of the sentence. However, only focusing on the features of target text is the main solution of most existing research. Moreover, the scale of publicly available Chinese sarcasm dataset is very small and does not contain the contextual information. To address the issues mentioned above, we build a Chinese sarcasm dataset from Bilibili, which is one of the most widely used social network platforms in China and has a significant number of sarcastic comments and contextual information. As far as we know, our dataset is the first publicly available large-scale Chinese sarcasm dataset including contextual information. Additionally, we have proposed a novel retrospective reading method for detecting sarcasm that leverages contextual information to improve model’s performance. The experimental results show the effectiveness of the proposed model and the significance of contextual information for Chinese sarcasm detection: achieving the highest F-score of 0.6942, outperforming existing state-of-the-art (SOTA) approaches. The study presented in this article offers approaches and ideas for future Chinese sarcasm detection studies.