Active learning (AL) attempts to maximize a model's performance gain while annotating the fewest samples possible. Deep learning (DL) is greedy for data and requires a large amount of data supply to optimize a massive number of parameters if the model is to learn how to extract high-quality features. In recent years, due to the rapid development of internet technology, we have entered an era of information abundance characterized by massive amounts of available data. As a result, DL has attracted significant attention from researchers and has been rapidly developed. Compared with DL, however, researchers have relatively low interest in AL. This is mainly because before the rise of DL, traditional machine learning requires relatively few labeled samples, meaning that early AL is rarely accorded the value it deserves. Although DL has made breakthroughs in various fields, most of this success is due to the large number of publicly available annotated datasets. However, the acquisition of a large number of high-quality annotated datasets consumes a lot of manpower, making it unfeasible in fields that require high levels of expertise (such as speech recognition, information extraction, medical images, etc.) Therefore, AL is gradually coming to receive the attention it is due.It is therefore natural to investigate whether AL can be used to reduce the cost of sample annotations, while retaining the powerful learning capabilities of DL. As a result of such investigations, deep active learning (DAL) has emerged. Although research on this topic is quite abundant, there has not yet been a comprehensive survey of DAL-related works; accordingly, this article aims to fill this gap. We provide a formal classification method for the existing work, along with a comprehensive and systematic overview. In addition, we also analyze and summarize the development of DAL from an application perspective. Finally, we discuss the confusion and problems associated with DAL and provide some possible development directions.CCS Concepts: • Computing methodologies → Machine learning algorithms.

No abstract

No abstract

Human cytochrome P450 oxidoreductase (POR) provides electrons for all microsomal cytochromes P450 (P450s) and plays an indispensable role in drug metabolism catalyzed by this family of enzymes. We evaluated 100 human liver samples and found that POR protein content varied 12.8-fold, from 12.59 to 160.97 pmol/mg, with a median value of 67.99 pmol/mg; POR mRNA expression varied by 26.4-fold. POR activity was less variable with a median value of 56.05 nmol/min per milligram. Cigarette smoking and alcohol consumption clearly influenced POR activity. Liver samples with a 2286822 TT genotype had significantly higher POR mRNA expression than samples with CT genotype. Homozygous carriers of POR2286822C>T, 2286823G>A, and 3823884A>C had significantly lower POR protein levels compared with the corresponding heterozygous carriers. Liver samples from individuals homozygous at 286823G>A, 1135612A>G, and 10954732G>A generally had lower POR activity levels than those from heterozygous or wild-type samples, whereas the common variant POR*28 significantly increased POR activity. There was a strong association between POR and the expression of P450 isoforms at the mRNA and protein level, whereas the relationship at the activity level, as well as the effect of POR protein content on P450 activity, was less pronounced. POR transcription was strongly correlated with both hepatocyte nuclear factor 4 alpha and pregnane X receptor mRNA levels. In conclusion, we have elucidated some potentially important correlations between POR single-nucleotide polymorphisms and POR expression in the Chinese population and have developed a database that correlates POR expression with the expression and activity of 10 P450s important in drug metabolism.

Hepatocellular carcinoma (HCC) is one of the most fatal cancers worldwide. Here, we show that expression of abnormal spindle-like microcephaly-associated (ASPM) is upregulated in liver cancer samples and this upregulation is significantly associated with tumor aggressiveness and reduced survival times of patients. Downregulation of ASPM expression inhibits the proliferation, invasion, migration and epithelial-to-mesenchymal transition of HCC cells in vitro, and inhibits tumor formation in nude mice. ASPM interacts with disheveled-2 (Dvl2) and antagonizes autophagy-mediated Dvl2 degradation by weakening the functional interaction between Dvl2 and the lipidated form of microtubule-associated proteins 1A/1B light chain 3A (LC3II), thereby increasing Dvl2 protein abundance and leading to Wnt/β-catenin signaling activation in HCC cells. Thus, our results define ASPM as a novel oncoprotein in HCC and indicate that disruption of the Wnt-ASPM-Dvl2-β-catenin signaling axis might have potential clinical value.