http://chineseinput.net/에서 pinyin(병음)방식으로 중국어를 변환할 수 있습니다.
변환된 중국어를 복사하여 사용하시면 됩니다.
Shu Zhang,Mei-qing Qiu,Hui-jun Wang,Ya-fei Ju,Zhen Liu,Tao Wang,Shi-feng Kan,Zhen Yang,Ya-yun Cui,You-qiang Ke,Hong-min He,Li Sun 대한위암학회 2023 Journal of gastric cancer Vol.23 No.2
Purpose: Gastric cancer (GC) is the second most lethal cancer globally and is associated with poor prognosis. Fatty acid-binding proteins (FABPs) can regulate biological properties of carcinoma cells. FABP5 is overexpressed in many types of cancers; however, the role and mechanisms of action of FABP5 in GC remain unclear. In this study, we aimed to evaluate the clinical and biological functions of FABP5 in GC. Materials and Methods: We assessed FABP5 expression using immunohistochemical analysis in 79 patients with GC and evaluated its biological functions following in vitro and in vivo ectopic expression. FABP5 targets relevant to GC progression were determined using RNA sequencing (RNA-seq). Results: Elevated FABP5 expression was closely associated with poor outcomes, and ectopic expression of FABP5 promoted proliferation, invasion, migration, and carcinogenicity of GC cells, thus suggesting its potential tumor-promoting role in GC. Additionally, RNA-seq analysis indicated that FABP5 activates immune-related pathways, including cytokine-cytokine receptor interaction pathways, interleukin-17 signaling, and tumor necrosis factor signaling, suggesting an important rationale for the possible development of therapies that combine FABP5-targeted drugs with immunotherapeutics. Conclusions: These findings highlight the biological mechanisms and clinical implications of FABP5 in GC and suggest its potential as an adverse prognostic factor and/or therapeutic target.
Bagging deep convolutional autoencoders trained with a mixture of real data and GAN-generated data
( Cong Hu ),( Xiao-jun Wu ),( Zhen-qiu Shu ) 한국인터넷정보학회 2019 KSII Transactions on Internet and Information Syst Vol.13 No.11
While deep neural networks have achieved remarkable performance in representation learning, a huge amount of labeled training data are usually required by supervised deep models such as convolutional neural networks. In this paper, we propose a new representation learning method, namely generative adversarial networks (GAN) based bagging deep convolutional autoencoders (GAN-BDCAE), which can map data to diverse hierarchical representations in an unsupervised fashion. To boost the size of training data, to train deep model and to aggregate diverse learning machines are the three principal avenues towards increasing the capabilities of representation learning of neural networks. We focus on combining those three techniques. To this aim, we adopt GAN for realistic unlabeled sample generation and bagging deep convolutional autoencoders (BDCAE) for robust feature learning. The proposed method improves the discriminative ability of learned feature embedding for solving subsequent pattern recognition problems. We evaluate our approach on three standard benchmarks and demonstrate the superiority of the proposed method compared to traditional unsupervised learning methods.