基于斯蒂芬森价值迭代的改进DDPG算法

针对DDPG算法的值函数迭代学习收敛缓慢以及经验利用率低的问题,提出一种基于Steffensen价值迭代和注意力经验回放的DDPG算法。将Steffensen迭代法应用于价值迭代过程,提高其收敛速度;采用基于注意力的经验回放机制,计算智能体当前所处状态与经验中状态的相似性,优先采样相似度高的经验,使智能体专注学习包含当前策略频繁访问的状态的经验。在PyBullet环境的6个连续动作控制任务中进行实验,其结果表明,相比DDPG算法、TD3算法、CDLR-DDPG算法和EER-DDPG算法,所提算法收敛更快,性能更好。

疫情背景下医学影像学远程在线教学的实践

探索新型冠状病毒肺炎疫情防控期间医学影像学远程在线教学模式是在保障学生生命健康的前提下完成的教学任务、保证教学质量的重要手段。通过总结学科课程在线教学实践,寻求切实可行的教学管理体制机制,运用网络智慧教学平台,落实教学要求,全面有序开展教学活动,以期实现医学影像学远程在线教学的全流程管控和良好教学效果的切实体现。

基于遗传算法优化单类支持向量机的油田离心泵注水站异常检测

目前油田离心泵注水站多采用传统的人工巡检等方式进行异常检测,存在浪费大量资源且检测精度不高的情况。针对此问题,提出一种基于遗传算法优化单类支持向量机(genetic algorithm optimized one-class support vector machine,GA-OC-SVM)的注水站异常检测方法。首先,对注水站数据进行标准化、归一化处理以及特征提取;其次,使用遗传算法进行寻优得到最佳种群个体值作为单类支持向量机的参数,建立检测模型;最后,将GA-OC-SVM算法与孤立森林算法、局部离群因子算法等主流方法用于测试数据集的异常检测对比,并分析算法的精度,采用接受者操作特征(receiver operating characteristics,ROC)曲线进行模型评价。结果表明所提出的GA-OC-SVM算法更优,检测精度达到99%,同时能够节省大量的人力物力资源。

化痰祛瘀中药联合溴隐亭含药血清抑制泌乳素垂体瘤细胞增殖的体外研究

目的探索化痰祛瘀中药联合溴隐亭含药血清对体外培养泌乳素垂体瘤细胞增殖的影响及其作用机制。方法以大鼠垂体瘤细胞系MMQ作为研究对象。将细胞分为空白对照组、空白血清组、溴隐亭组(溴隐亭0.104 mg)、联合治疗低剂量(化痰祛瘀中药6.25 ml+溴隐亭0.104 mg)、中剂量(化痰祛瘀中药12.5 ml+溴隐亭0.104 mg)及高剂量组(化痰祛瘀中药25 ml+溴隐亭0.104 mg),给予相应干预措施,通过MTT法检测不同干预措施对MMQ细胞增殖的影响。采用流式细胞仪检测细胞周期,免疫印迹法检测多巴胺受体DRD2及抑癌基因PETN蛋白表达。结果干预24 h后,溴隐亭组细胞存活率为77.83%,联合治疗低、中及高剂量组细胞存活率分别为74.64%、66.67%和58.69%;其中溴隐亭组、联合治疗中及高剂量组细胞存活率显著低于空白血清组(P<0.05或P<0.01)。干预48 h后,溴隐亭组细胞存活率为59.44%,联合治疗低、中及高剂量组细胞存活率分别为57.62%、54.41%和45.87%,且溴隐亭组及联合治疗各剂量组细胞存活率均显著低于空白血清组(P<0.05或P<0.01);同时,联合治疗高剂量组细胞存活率显著低于溴隐亭组(P<0.05)。干预72 h后,溴隐亭组细胞存活率为49.56%,联合治疗低、中及高剂量组细胞存活率分别为46.65%、30.01%和23.78%,且溴隐亭组及联合治疗各剂量组细胞存活率均显著低于空白血清组(P<0.05或P<0.01);联合治疗中及高剂量组细胞存活率低于溴隐亭组(P<0.05或P<0.01)。溴隐亭组及联合治疗各剂量组细胞存活率随着时间的延长而降低(P<0.05)。联合治疗高剂量组细胞周期增殖S期比例显著降低,从25.61%下调为5.83%(P<0.01)。此外,化痰祛瘀中药联合溴隐亭治疗可上调细胞多巴胺受体及抑癌基因PTEN蛋白表达。结论化痰祛瘀中药通过上调多巴胺受体蛋白表达,可能是其协同增效溴隐亭治疗泌乳素垂体瘤,发挥临床治疗作用的主要途径。

Regulatory Effect of Electroacupuncture on Ceramide Galactosyltransferase Expression in Sciatic Nerve of Experimental Diabetic Rats

Objective: To observe the change of ceramide galactosyltransferase(CGT) expression in sciatic nerve of experimental diabetic rats after electroacupuncture(EA) treatment. Methods: The rat model of diabetic peripheral neuropathy(DPN) was established with Streptozotocin(STZ). Twenty-eight rats were randomly divided into a normal group, a model group, an EA acupoint group, and an EA non-acupoint group, with 7 rats in each group. Acupuncture on points Shenshu(BL 23) and Zusanli(ST 36) was performed in the EA group using EA electronic device with continuous wave, 2 Hz, 20 min every time, once every other day for 12 times. Acupuncture(on the tip of rat tail) in EA non-acupoint group was performed with the same electronic parameters and time. The model group and the normal group were not given any treatment, except the same fixation as that in the EA acupoint group. After 12 treatments, the albumen and mRNA expressions of CGT of sciatic nerve in rats from each group were measured with real-time polymerase chain reaction(RT-PCR), fluorescence quantitative PCR(FQ-PCR) and immunohistochemical methods. Results: After modeling, the CGT expression was significantly lower than that in the normal group(P<0.01); however, after 12 EA treatments, the CGT positive cells in the EA acupoint group were up-regulated and significantly higher than those in the EA non-acupoint group and the model group(P<0.01). The CGT mRNA expression of the rats after modeling was also higher than that of the normal rats(P<0.01), and the expression in the EA acupoint group was significantly lower than those in the model group and the EA non-acupoint group(P<0.01, P<0.05). Conclusion: EA treatment with acupoints displays some therapeutic or preventive effect in the recovery of DPN via reversing the abnormal expression of CGT mRNA in sciatic nerves of experimental diabetic neuropathy rats, and it is worthy of further study.