融合Dead-ends和离线监督Actor-Critic的动态治疗策略生成模型

强化学习对数学模型依赖性低,利用经验便于架构和优化模型,非常适合用于动态治疗策略学习。但现有研究仍存在以下问题:1)学习策略最优性的同时未考虑风险,导致学到的策略存在一定的风险;2)忽略了分布偏移问题,导致学到的策略与医生策略完全不同;3)忽略患者的历史观测数据和治疗史,从而不能很好地得到患者状态,进而导致不能学到最优策略。基于此,提出了融合Dead-ends和离线监督Actor-Critic的动态治疗策略生成模型DOSAC-DTR。首先,考虑学到的策略所推荐的治疗行动的风险性,在Actor-Critic框架中融入Dead-ends概念;其次,为缓解分布偏移问题,在Actor-Critic框架中融入医生监督,在最大化预期回报的同时,最小化所学策略与医生策略之间的差距;最后,为了得到包含患者关键历史信息的状态表示,使用基于LSTM的编码器解码器模型对患者的历史观测数据和治疗史进行建模。实验结果表明,DOSAC-DTR相比基线方法有更好的性能,可以得到更低的估计死亡率以及更高的Jaccard系数。

On the Problem of All Equal Positive Odd Number Solutions of the Dissemination of the Ends' Guess

The thesis is concerned with the problem of the dissemination of the Ends’ guess and offers the formula of all equal positive odd number solutions.

New Cutting Force Modeling Approach for Flat End Mill

A new mechanistic cutting force model for flat end milling using the instantaneous cutting force coefficients is proposed. An in-depth analysis shows that the total cutting forces can be separated into two terms: a nominal component independent of the runout and a perturbation component induced by the runout. The instantaneous value of the nominal component is used to calibrate the cutting force coefficients. With the help of the perturbation component and the cutting force coefficients obtained above, the cutter runout is identified. Based on simulation and experimental results, the validity of the identification approach is demonstrated. The advantage of the proposed method lies in that the calibration performed with data of one cutting test under a specific regime can be applied for a great range of cutting conditions.

Observations of boundary layer parameters and suspended sediment transport over the intertidal flats of northern Jiangsu, China

A current-turbidity monitoring system (CTMS) was deployed on the intertidal flat at Wanggang, northern Jiangsu during October 16-17, 2000, to measure the tidal current speeds and seawater turbidities at 5 levels above the seabed. Based upon the logarithmic-profile equation, the boundary layer parameters, i.e., u, z0 and C60, were obtained for 247 tidal flow velocity profiles. Around 90% of the profiles were logarithmic according to the critical correlation coefficient. Internal consistency analysis shows that these parameters derived by different methods are consistent with each other. In addition, the height of the bedforms observed is close to the seabed roughness lengths calculated from the velocity profiles, indicating that the boundary layer parameters obtained can reveal the conditions at the sediment-water interface on the intertidal flats. Suspended sediment concentrations were obtained from the 5 CTMS turbidity meters using laboratory and in-situ calibrations. The results show that the in-situ calibrated SSCs have a much higher accuracy than the laboratory calibrated ones. Calculation of suspended sediment fluxes on the intertidal flats, with a magnitude of 104 kg/m per spring tidal cycle, indicates that suspended sediment moves towards the northwest, which is reversal to the transport pattern controlled by the southward Northern Jiangsu Coastal Current in the sub-tidal zone and adjacent shallow waters.

New Mathematical Method for the Determination of Cutter Runout Parameters in Flat-end Milling

The cutting force prediction is essential to optimize the process parameters of machining such as feed rate optimization, etc. Due to the significant influences of the runout effect on cutting force variation in milling process, it is necessary to incorporate the cutter runout parameters into the prediction model of cutting forces. However, the determination of cutter runout parameters is still a challenge task until now. In this paper, cutting process geometry models, such as uncut chip thickness and pitch angle, are established based on the true trajectory of the cutting edge considering the cutter runout effect. A new algorithm is then presented to compute the cutter runout parameters for flat-end mill utilizing the sampled data of cutting forces and derived process geometry parameters. Further, three-axis and five-axis milling experiments were conducted on a machining centre, and resulting cutting forces were sampled by a three-component dynamometer. After computing the corresponding cutter runout parameters, cutter forces are simulated embracing the cutter runout parameters obtained from the proposed algorithm. The predicted cutting forces show good agreements with the sampled data both in magnitude and shape, which validates the feasibility and effectivity of the proposed new algorithm of determining cutter runout parameters and the new way to accurately predict cutting forces. The proposed method for computing the cutter runout parameters provides the significant references for the cutting force prediction in the cutting process.

Optimization of cutter-location for 5-axis NC machining freeform surface with a flat-end cutter

axis NC machining freeform surface with a flat end cutter can theoretically improve material removal rate and surface finish, but it is very difficult to generate a gouge free cutter location. A new method of generating optimal cutter path is proposed to define the effective radius of a flat end cutter and determine the optimal step forward distance and step over distance. Thereby improving the NC machining efficiency and quality of freeform surfaces.

Comparative study of mining methods for reserves beneath end slope in flat surface mines with ultra-thick coal seams

The paper aims to identify a reasonable method for mining ultra-thick coal seams in an end-slope in surface mine, With a case study of Heidaigou surface coal mine(HSCM), the paper conducted a comparative research on three mining methods, namely Underground Mining Method(UMM), Highwall Mining System(HMS) and Local Steep Slope Mining Method(LSSMM). A model was firstly established to simulate the impact that UMM and HMS exert on monitoring points and surface deformation. The way that stripping and excavation amount varies with different slope angle, and the corresponding end slope stability were analyzed in the mode of LSSMM. Then a TOPSIS model was established by taking into account six indicators such as recovery ratio, technical complexity and adaptability, the impact on surface mining production, production safety and economic benefits. Finally, LSSMM was determined as the best mining method for mining ultra-thick coal seams in end slope in HSCM.

Electrostatic Actuating Bendable Flat Electrode for Micro Electrochemical Machining

In micro-electrochemical machining(μECM), material dissolution takes place at very close vicinity of tool electrode due to localization of electric field. Controlling the gap between tool electrode and workpiece is the key to μECM. Therefore, a new method is proposed to solve a variety of problems in small gap control. In the present context, experiments were carried out with an indigenously developed setup to fabricate cylindrical arrays. During the machining process, the flat electrode bends due to electrostatic force in pulse on-time, which self-adaptively narrows the gap between the electrode and the workpiece. The workpiece material will be removed once the gap meets the processing condition. Therefore, this method has advantages of reducing dependence on high precision machine tools and of avoiding complex servo control. The flat electrode quickly restores to its original condition when it is in pulse off-time, making the gap much larger than that in traditional electrochemical machining(ECM). The large gap benefits debris removing, which improves the machining accuracy. The influence of different experimental parameters on accuracy and efficiency during the machining process has been investigated. It is observed that with the increase in applied voltage or concentration of electrolyte, the material removal rate and the process gap both increase. The detailed analysis of the experimental results is described in this paper.

Large magnetic moment at sheared ends of single-walled carbon nanotubes

In this work we report that after single-walled carbon nanotubes(SWNTs) are sheared with a pair of titanium scissors,the magnetization becomes larger than that of the corresponding pristine ones. The magnetization increases proportionally with the number of SWNTs with sheared ends, suggesting that there exist magnetic moments at the sheared ends of SWNTs.By using the coefficient of this linear relation, the average magnetic moment is estimated to be 41.5 ± 9.8 μB(Bohr magneton) per carbon atom in the edge state at temperature of 300.0 K, suggesting that ultrahigh magnetic fields can be produced. The dangling sigma and pi bonds of the carbon atoms at sheared ends play important roles in determining the unexpectedly high magnetic moments, which may have great potential applications.

Giant magnetic moment at open ends of multiwalled carbon nanotubes

The attractions of cantilevers made of multiwalled carbon nanotubes（MWNTs） and secured on one end are studied in the non-uniform magnetic field of a permanent magnet. Under an optical microscope, the positions and the corresponding deflections of the original cantilevers（with iron catalytic nanoparticles at the free end） and corresponding cut-off cantilevers（the free ends consisting of open ends of MWNTs） are studied. Both kinds of CNT cantilevers are found to be attracted by the magnet, and the point of application of force is proven to be at the tip of the cantilever. By measuring and comparing deflections between these two kinds of cantilevers, the magnetic moment at the open ends of the CNTs can be quantified.Due to the unexpectedly high value of the magnetic moment at the open ends of carbon nanotubes, it is called giant magnetic moment, and its possible mechanisms are proposed and discussed.

Strain Transfer Mechanism of Grating Ends Fiber Bragg Grating for Structural Health Monitoring

The grating ends bonding fiber Bragg grating(FBG)sensor has been widely used in sensor packages such as substrate type and clamp type for health monitoring of large structures.However,owing to the shear deformation of the adhesive layer of FBG,the strain measured by FBG is often different from the strain of actual matrix,which causes strain measurement errors.This investigation aims at improving the measurement accuracy of strain for the grating ends surface-bonded FBG.To fulfill this objective,a strain transfer equation of the grating ends bonding FBG is derived,and a theoretical model of the average strain transfer from the matrix to the optical fiber is developed.Moreover,parameters that influence the average strain transfer rate from the matrix to the optical fiber are analyzed.A selection scheme of bonding parameters by numerical simulation is provided,which is significantly advantageous over that of the grating bonding FBG.The theoretical equation is verified by finite element method(FEM).Compared with the existing model,the proposed model has higher measurement accuracy.Experimental tests are performed to validate the effectiveness of the proposed model on the equalintensity cantilever beam,whose surface is attached to the bare FBG with grating ends bonding and strain gauge by using epoxy glue.The results show that there is a great agreement between the outcome of the bare FBG and that of the strain gauge,and the corrected strain is closer to the true strain.The proposed model provides a theoretical basis for the design of the grating ends surface-bonded FBG strain sensor for health monitoring of large structures.

The hepatitis C virus 5’ untranslated region gene amplified by rapid amplification of cDNA ends and its secondary structure

Objectives: To obtain very end full-length cDNA ofhepatitis C virus (HCV) 5’ untranslated region(5’UTR) and analyze its primary and secondarystructure.Methods: A patient infected genotype 2a HCV wasidentified by reverse transcription-nested polymerasechain reaction (RT-PCR) and restriction fragmentlength polymorphism (RFLP). Total RNA isolatedfrom the serum was used as template, and the cDNAof the 5’ untranslated region was amplified using rap-id amplification of cDNA ends (RACE). The frag-ments were recombinated by A-T clone strategy, andthe recombinants were confirmed by RFLP andPCR, and sequenced subsequently. Secondary struc-tures were analysed by RNAdraw.Results: Very end full-length cDNA of genotype 2aHCV 5’ UTR was obtained by RACE. In five clonesobtained, three contained full-length 5’UTR cDNA;A21G, G170A, T222C, T247C, C339T substitutionswere found as compared to HC-J6. Homological re-sults of HCV-1, HC-J6, HC-C2, HC-J8 were 93.6%-94.4%, 92.1%-93%, 98.8%-99.7%, 96.2%-96.5%, respectively; however, the substitutions didnot alter secondary structure. Two of 5 clones weredeletions of 53bp and 135bp at the 5’terminal ofHCV 5’UTR, respectively.Conclusions: RACE can be used to obtain the full-length cDNA of 2a genotype HCV 5’UTR. Genes de-leted at the 5’ terminal of HCV circulate in hepatitisC patients.

Suitable retention and recovery technology of floor coal at ends of fully mechanized face with great mining heights

Gateways at faces of great mining heights are mostly driven along the roof of coal seams.For gateway height restrictions,a 1-3 m floor coal is retained,leaving a triangular floor coal at the face ends,causing a loss of coal.In order to improve coal recovery rates and to ensure efficiency of equipment at coal mining faces,we investigated suitable retention methods and recovery technology of floor coal at face ends.The upper floor coal can directly be recovered by a shearer with floor dinting.The lower floor coal is recovered by shearer with floor dinting after advanced floor dinting and retaining a step for protecting coal sides in a haulage gateway.Field practice shows that this method can improve the coal recovery rates at fully mechanized working faces with great mining heights.

National trends in resection of the distal pancreas

AIM:To investigate national trends in distal pancreatectomy(DP) through query of three national patient care databases.METHODS:From the Nationwide Inpatient Sample(NIS,2003-2009),the National Surgical Quality Improvement Project(NSQIP,2005-2010),and the Surveillance Epidemiology and End Results(SEER,2003-2009) databases using appropriate diagnostic and procedural codes we identified all patients with a diagnosis of a benign or malignant lesion of the body and/or tail of the pancreas that had undergone a partial or distal pancreatectomy.Utilization of laparoscopy was defined in NIS by the International Classification of Diseases,Ninth Revision correspondent procedure code;and in NSQIP by the exploratory laparoscopy or unlisted procedure current procedural terminology codes.In SEER,patients were identified by the International Classification of Diseases for Oncology,Third Edition diagnosis codes and the SEER Program Code Manual,third edition procedure codes.We analyzed the databases with respect to trends of inpatient outcome metrics,oncologic outcomes,and hospital volumes in patients with lesions of the neck and body of the pancreas that underwent operative resection.RESULTS:NIS,NSQIP and SEER identified 4242,2681 and 11 082 DP resections,respectively.Overall,laparoscopy was utilized in 15%(NIS) and 27%(NSQIP).No significant increase was seen over the course of the study.Resection was performed for malignancy in 59%(NIS) and 66%(NSQIP).Neither patient Body mass index nor comorbidities were associated with operative approach(P = 0.95 and P = 0.96,respectively).Mortality(3% vs 2%,P = 0.05) and reoperation(4% vs 4%,P = 1.0) was not different between laparoscopy and open groups.Overall complications(10% vs 15%,P < 0.001),hospital costs [44 741 dollars,interquartile range(IQR) 28 347-74 114 dollars vs 49 792 dollars,IQR 13 299-73 463,P = 0.02] and hospital length of stay(7 d,IQR 4-11 d vs 7 d,IQR 6-10,P < 0.001) were less when laparoscopy was utilized.One and two year survival after resection for malignancy were unchanged over the course of the study(ductal adenocarinoma 1-year 63.6% and 2-year 35.1%,P = 0.53;intraductal papillary mucinous neoplasm and nueroendocrine 1-year 90% and 2-year 84%,P = 0.25).The majority of resections were performed in teaching hospitals(77% NIS and 85% NSQIP),but minimally invasive surgery(MIS) was not more likely to be used in teaching hospitals(15% vs 14%,P = 0.26).Hospitals in the top decile for volume were more likely to be teaching hospitals than lower volume deciles(88% vs 43%,P < 0.001),but were no more likely to utilize MIS at resection.Complication rate in teaching and the top decile hospitals was not significantly decreased when compared to non-teaching(15% vs 14%,P = 0.72) and lower volume hospitals(14% vs 15%,P = 0.99).No difference was seen in the median number of lymph nodes and lymph node ratio in N1 disease when compared by year(P = 0.17 and P = 0.96,respectively).CONCLUSION:There appears to be an overall underutilization of laparoscopy for DP.Centralization does not appear to be occurring.Survival and lymph node harvest have not changed.

Korea Peninsula Landscape Extends from Continent End

The Korea peninsula extends southward from the eastern end of the Asian continent. It is divided just slightly north of the 38th parallel. The peninsula is roughly 1020 km(612 miles) long and 175 km( 105 miles) wide at its narrowest point.

基于情感映射的动漫形象价值观诉求分析——以LINE FRIENDS为例

"LINE FRIENDS"系列产品的深入人心,离不开其扁平化形象所呈现的简单生动和营销策略的成功。本文以LINE FRIENDS为例,从情感化设计入手分析动漫形象的主要特征所折射的性格特点,和在视觉符号与触觉符号的共同作用下给用户留下的深刻记忆,得出动漫形象情感化会让这一形象更深入人心,引发共鸣。

Endoscopic submucosal dissection of multiple flat adenomas in the radiated rectum

We report a case of multiple flat adenomas and cancer of the rectum that occurred 15 years after pelvic irradiation following surgery for uterine cancer.Adenoma borders were diagnosed accurately by magnifying chromoendoscopy,leading to their adequate excision using endoscopic submucosal dissection.This enabled minimal dissection of the irradiated pelvis that would have otherwise been difficult.Furthermore,our approach probably helped minimize loss of bowel function,thereby preserving the patient's quality of life as much as possible.Pathology of the resected specimens revealed thickened walls of the submucosal layer vessels,indicating chronic radiation proctitis.Pelvic irradiation of the bowel carries a high risk of causing flat adenomas and cancer.Close and long-term surveillance may be useful in such cases,using not only conventional colonoscopy but also chromoendoscopy with indigo carmine dye spray and magnifying endoscopy.

Trends of Shoreline Position: An Approach to Future Prediction for Balasore Shoreline, Odisha, India

The present study aims to analyze the shift in shoreline due to coastal processes and formulate available for best estimate of future shoreline positions based on precedent shorelines. Information on rates and trends of shoreline change can be used to improve the understanding of the underlying causes and potential effects of coastal erosion which can support informed coastal management decisions. In this paper, researchers go over the changes in the recent positions of the shoreline of the Balasore coast for the 38 years from 1975 through 2013. The study area includes the Balasore coastal region from Rasalpur to Udaypur together with Chandipur, Choumukh, Chandrabali as well as Bichitrapur. Transects wise shoreline data base were developed for approximately 67 kilometers of shoreline and erosional/accretional scenario has also been analysed by delineating the shoreline from Landsat imageries of 1975, 1980, 1990, 1995, 2000, 2005, 2010 and 2013. A simple Linear Regression Model and End Point Rate (EPR) have been adopted to take out the rate of change of shoreline and its future positions, based on empirical observations at 67 transects along the Balasore coast. It is found that the north eastern part of Balasore coast in the vicinity of Subarnarekha estuary and Chandrabali beach undergo high rates of shore line shift. The shoreline data were integrated for long- (about 17 years) and short-term (about 7 years) shift rates analysis to comprehend the shoreline change and prediction. For the prediction of future shoreline, the model has been validated with the present shoreline position (2013). The rate of shoreline movement calculated from the fixed base line to shoreline position of 1975, 1980, 1990, 1995, 2000, 2005 and 2010 and based on this, the estimated shoreline of 2013 was calculated. The estimated shoreline was compared with the actual shoreline delineated from satellite imagery of 2013. The model error or positional shift at each sample point is observed. The positional error varies from??4.82 m to 212.41 m. It has been found that model prediction error is higher in the left hand side of river Subarnarekha. The overall error for the entire predicted shoreline was found to be 41.88 m by Root Mean Square Error (RMSE). In addition, it was tested by means difference between actual and predicted shoreline positions using “t” test and it has been found that predicted shore line is not significantly different from actual shoreline position at (t132 = 0.278) p < 0.01.

Influencing Factors of the Power Fluctuation on the Ultra High Voltage Transmission Line Caused by Faults at the Remote Ends of the Interconnected Grid

After the North China grid and the Central China grid get into connection with the UHVAC demonstration, a new phenomenon is discovered according to some simulations. That is, the faults at the remote end of the UHV interconnected grid will result in significant power fluctuation and voltage drop on the UHV transmission line and even system splitting. But the faults near the UHV line only have marginal effects. Further, the simulation results also indicate that the short-circuit current of the buses near the UHV line is larger than that of the buses far away from the UHV line. This phenomenon is divergent from the traditional view. In this paper, the detail will be introduced, and the factors influencing the system stability after faults are presented and analyzed. The results indicate that transmission power of the UHV line and of the lines between the remote end and the major grid influence the fluctuation on UHV line. The load model and the grid structure of the remote end also have effect on it. Finally, corresponding control scheme is presented to improve the operation conditions of the UHV interconnected grid and ensure its security and stability.

FE Study for Reducing Forming Forces and Flat End Areas of Cylindrical Shapes Obtained by the Roll-Bending Process

A roll-bending process that minimizes the flat areas on the leading and trailing ends of formed plates will produce more accurate and easier assemble final shapes. There are several methods of minimizing flat areas, but they are costly or difficult to apply for thick plates. This study proposes a new, simple approach that reduces these flat areas. This approach includes moving the bottom roll slightly along the feeding direction and adjusting the bottom roll location. Sensitivity analyses were performed using a developed 3D dynamic FE （finite element） model of an asymmetrical roll-bending process in the Ansys/LS-Dyna software package. Simulations were validated by experiments run on an instrumented roll-bending machine. The FE results indicate that this new approach not only minimizes the flat areas but also reduces the forming forces.