Deep Neural Networks (DNNs) have achieved tremendous success in various tasks. However, DNNs exhibit uncertainty and unreliability when faced with well-designed adversarial examples, leading to misclassification. To address this, a variety of methods have been proposed to improve the robustness of DNNs by detecting adversarial attacks. In this paper, we combine model explanation techniques with adversarial models to enhance adversarial detection in real-world scenarios. Specifically, we develop a novel adversary-resistant detection framework called EXPLAINER, which utilizes explanation results extracted from explainable learning models. The explanation model in EXPLAINER generates an explanation map that identifies the relevance of input variables to the model’s classification result. Consequently, adversarial examples can be effectively detected by comparing the explanation results of a given sample with its denoised version, without relying on any prior knowledge of attacks. The proposed framework is thoroughly evaluated against different adversarial attacks, and experimental results demonstrate that our approach achieves promising results in white-box attack scenarios.

Finding relationships between two related groups of images has been a challenging question in many fields. E.g., it is informative in neuron science studies to build the connection between experiment animals' neuron activity images and their behavior images. Very few previous works have achieved this task since most generative models focus on reconstructing the output images similar to input images. We proposed a novel framework in this paper to accomplish this goal, which could map images from one group to images from another group. We apply the singular value decomposition (SVD) method to remove the original images' background noise. Next, we combine two deep learning approaches, variational autoencoder (VAE) and convolutional neural networks (CNN), to directly connect two groups of images. We test our framework on images from a neuron science experiment. Results show that the proposed framework could generate mice paw movement images given the mice neuron images, which are very close to the ground truth images. In terms of capturing the paw gestures in paw movement images, experiment results demonstrate that our framework outperforms the state-of-art paw location detection method.

Since the first identified case of COVID-19 in December 2019, a plethora of pharmaceuticals and therapeutics have been tested for COVID-19 treatment. While medical advancements and breakthroughs are well underway, the sheer number of studies, treatments, and associated reports makes it extremely challenging to keep track of the rapidly growing COVID-19 research landscape. While existing scientific literature search systems provide basic document retrieval, they fundamentally lack the ability to explore data, and in addition, do not help develop a deeper understanding of COVID-19 related clinical experiments and findings. As research expands, results do so as well, resulting in a position that is complicated and overwhelming. To address this issue, we present a named entity recognition based framework that accurately extracts COVID-19 related information from clinical test results articles, and generates an efficient and interactive visual knowledge graph. This knowledge graph platform is user friendly, and provides intuitive and convenient tools to explore and analyze COVID-19 research data and results including medicinal performances, side effects and target populations.

Since the first cased of COVID-19 was identified in December 2019, a plethora of different drugs have been tested for COVID-19 treatment, making it a daunting task to keep track of the rapid growth of COVID-19 research landscape. Using the existing scientific literature search systems to develop a deeper understanding of COVID-19 related clinical experiments and results turns to be increasingly complicated. In this paper, we build a named entity recognition-based framework to extract information accurately and generate knowledge graph efficiently from a myriad of clinical test results articles. Of the tested drugs to treat COVID-19, we also develop a question answering system answers to medical questions regarding COVID-19 related symptoms using Wikipedia articles. We combine the state-of-the-art question answering model - Bidirectional Encoder Representations from Transformers (BERT), with Knowledge Graph to answer patients' questions about treatment options for their symptoms. This generated knowledge graph is user-friendly with intuitive and convenient tools to find the supporting and/or contradictory references of certain drugs with properties such as side effects, target population, etc. The trained question answering platform provides a straightforward and error-tolerant way to query for treatment suggestions given uses' input symptoms.