http://chineseinput.net/에서 pinyin(병음)방식으로 중국어를 변환할 수 있습니다.
변환된 중국어를 복사하여 사용하시면 됩니다.
Artificial Intelligence in Transmission Control Clutch Engagement with Reinforcement Learning
Alexander Lampe,Clemens Guhmann,Roland Serway,Leopold Groβe Siestrup 한국자동차공학회 2021 한국자동차공학회 부문종합 학술대회 Vol.2021 No.6
Software development for transmissions mainly uses model-based control concepts that require physical modeling of the system. The drawback is that knowledge of the real system is needed, which can usually only be represented with limited quality, resulting in augmented, time-consuming calibration effort for fine-tuning of the system. Transmission software functions such as the start-up controller for automatic transmissions with friction clutch as start-up element playa key role in defining system comfort and thus require a particularly high calibration effort. At the same time, non-linear system phenomena and uncertain control variables make it difficult to design a start-up controller that meets customer demands. In order to reduce calibration effort on the one hand and improve control quality on the other, a data-based controller has been developed with the reinforcement learning approach (RL). The advantages compared to model-based and model-predictive approaches such as model predictive control (MPC) result from the fact that is not strictly necessary to model the dynamic system (model-free RL) and the RL algorithm works on the basis of experience. The controller behavior is derived from minimizing the cost function and maximizing the reward function. In other words, the reinforcement learning controller operates optimally in terms of a chosen cost function. This paper presents and discusses reinforcement learning as a machine learning process in the context of its use in automotive embedded systems, looking among others at the applicability of the method with regard to complying with functional safety and in terms of hardware requirements. The paper also gives an insight into the selection criteria for the algorithm and the procedure for stipulating important variables. Finally, initial simulation results are presented based on a simple powertrain model.
Diet and Cancer Prevention Research: From Mechanism to Implementation
Johanna W. Lampe 대한암예방학회 2020 Journal of cancer prevention Vol.25 No.2
Between 30% to 50% of cancer cases are estimated to be preventable through reduced exposure to tobacco, occupational carcinogens, and infectious agents, and adoption of lifelong healthy eating and a physically active lifestyle. In the past, diet and cancer prevention research has aimed to understand the effects of specific foods and nutrients on cancer-related mechanisms. More recently, there has been a shift in emphasis toward a more holistic focus on patterns of diet, reflecting the goal to understand the impact of adhering to broader public health recommendations. It is increasingly apparent from observational studies that different patterns of diet and physical activity are manifest in a metabolic state that is more, or less, conducive to the acquisition of genetic and epigenetic alterations leading to carcinogenesis. Experimental studies in cell systems, animals and humans have expanded our understanding of the many mechanisms by which specific dietary constituents may modulate inflammation and immune function, carcinogen metabolism, hormone and growth-factor regulation, DNA repair capacity, cell-cycle control, and proliferation and apoptosis. However, few mechanistic studies in animal models have evaluated diets containing the complex mixtures that make up human diets. Overall, more studies are needed across the continuum of prevention research, from basic mechanistic research on the effects of diet patterns on fundamental biologic processes to studies testing the efficacy of implementing lifestyle-directed cancer prevention strategies.
Modulation of Biotransformation Enzymes by Phytochemicals : Impact of Genotypes
Johanna W. Lampe 한국식품영양과학회 2004 한국식품영양과학회 학술대회발표집 Vol.55 No.-
Modulation of biotransformation enzymes is one mechanism by which a diet high in fruits and vegetables may influence cancer risk. Inhibition of cytochrome P450s (CYP) and concomitant induction of conjugating enzymes are hypothesized to reduce the impact of carcinogens in humans. Thus, exposure to types and amounts of phytochemicals may influence disease risk. Like other xenobiotics, many classes of phytochemicals are rapidly conjugated with glutathione, glucuronide, and sulfate moieties and excreted in urine and bile. In humans, circulating phytochemical levels vary widely among individuals even in response to controlled dietary interventions. Polymorphisms in biotransformation enzymes, such as the glutathione S-transferases (GST), UDP-glucuronosyltransferases (UGT), and sulfotransferases (SULT), may contribute to the variability in phytochemical clearance and efficacy; polymorphic enzymes with lower enzyme activity prolong the half-lives of phytochemicals in vivo. Isothiocyanates (ITC) in cruciferous vegetables are catalyzed by the four major human GSTs; however reaction velocities of the enzymes differ greatly. In some observational studies of cancer, polymorphisms in the GSTM1 and GSTT1 genes that result in complete lack of GSTM1-1 and GSTT1-1 protein, respectively, confer greater protection from cruciferous vegetables in individuals with these genotypes. Similarly, we have shown in a controlled dietary trial that levels of GST-alpha-induced by ITC-are higher in GSTM1-null individuals exposed to cruciferous vegetables. The selectivity of glucuronosyl conjugation of flavonoids is dependent both on flavonoid structure as well as on the UGT isozyme involved in its conjugation. The effects of UGT polymorphisms on flavonoid clearance have not been examined; but polymorphisms affect glucuronidation of several drugs. Given the strong interest in the chemopreventive effects of flavonoids, systematic evaluation of these polymorphic UGTs and flavonoid pharmacokinetics are warranted. Overall, these studies suggest that for phytochemicals that are metabolized by, and affect activity of, biotransformation enzymes, interactions between genetic polymorphisms in the enzymes and intake of the compounds should be considered in studies of cancer risk. Genetic polymorphisms in biotransformation enzymes may account in part for individual variation in metabolism of a wide range of phytochemicals and their ultimate impact on health.
Fink, Stephen P,Yang, Dong-Hoon,Barnholtz-Sloan, Jill S,Ryu, Yeon-Mi,Mikkola, Debra,Potter, John D,Lampe, Johanna W,Markowitz, Sanford D,Myung, Seung-Jae Plenum Pub. Corp.] 2013 Digestive diseases and sciences Vol.58 No.9
<P>15-Hydroxprostaglandin dehydrogenase (15-PGDH) mediates a colon neoplasia suppressor pathway, acting through metabolic antagonism of cyclooxygenase-mediated colon carcinogenesis. To determine whether the colon tumor prevention activity of 15-PGDH acts as a constant or variable effect among individuals, we determined whether 15-PGDH levels remain stable over subsite and time in the human colon, determined the extent of differences in 15-PGDH levels between different individuals, and determined whether 15-PGDH modulation mediates any part of the anti-colon tumor effect of aspirin.</P>
( Peter Takacs ),( Bence Kozma ),( Rudolf Lampé ),( Attila Sipos ),( Robert Poka ) 대한산부인과학회 2020 Obstetrics & Gynecology Science Vol.63 No.3
Objective To improve pelvic floor recovery after vaginal delivery with daily supplementation of a specially formulated postpartum recovery supplement. Methods Within 48 hours of vaginal delivery, primipara women were randomized in a 1:1 ratio to receive daily oral supplementation for 6 weeks with either a combination of regular prenatal vitamin (PNV), leucine (4 g/day), zinc (30 mg/day) and omega-3 fatty acid (900 mg/day) (treatment group), or only a PNV daily (control group). Co-primary outcomes were vaginal squeeze pressure as measured by perineometer and levator muscle injury as measured by transperineal 3-dimensional tomographic ultrasound at 6 weeks postpartum. Results Twenty-six women in the control group and 27 in the treatment group completed the trial. Weak pelvic floor muscle strength was significantly less frequent in the treatment group compared to the control group at 6 weeks after delivery (28% vs. 58%, P=0.03). Both right and left-sided levator-urethra gap was significantly larger in the control group compared to the treatment group indicating more levator injury being present in the control group at 6 weeks after delivery. Anterior vaginal wall prolapse at or beyond the hymenal ring was significantly more common in the control group compared to the treatment group (19% vs. 0%, P=0.02). Significantly more women reported bothersome bulge symptoms in the control group compared to the treatment group at 6 weeks postpartum (19% vs. 0%, P=0.02). Conclusion Postpartum women who received a specially formulated postpartum recovery supplement had improved recovery of the pelvic floor after vaginal delivery.