共找到 20 条结果
暂无摘要（点击查看原文获取完整内容）
Automatically generating coherent and semantically meaningful text has many applications in machine translation, dialogue systems, image captioning, etc. Recently, by combining with policy gradient, Generative Adversarial Nets(GAN) that use a discriminative model to guide the training of the generative model as a reinforcement learning policy has shown promising results in text generation. However, the scalar guiding signal is only available after the entire text has been generated and lacks intermediate information about text structure during the generative process. As such, it limits its success when the length of the generated text samples is long (more than 20 words). In this paper, we propose a new framework, called LeakGAN, to address the problem for long text generation. We allow the discriminative net to leak its own high-level extracted features to the generative net to further help the guidance. The generator incorporates such informative signals into all generation steps through an additional MANAGER module, which takes the extracted features of current generated words and outputs a latent vector to guide the WORKER module for next-word generation.Our extensive experiments on synthetic data and various real-world tasks with Turing test demonstrate that LeakGAN is highly effective in long text generation and also improves the performance in short text generation scenarios. More importantly, without any supervision, LeakGAN would be able to implicitly learn sentence structures only through the interaction between MANAGER and WORKER.
BACKGROUND: Esophageal anastomotic leak is a potentially life threatening complication of esophagectomy and esophagogastrectomy. We reviewed our experience with this complication and tried to identify factors predictive of mortality after esophageal anastomotic leak. METHODS: Records of patients undergoing esophagectomy and esophagogastrectomy for benign or malignant disease over a 10-year period (1989-1999), who developed esophageal anastomotic leaks, were reviewed. RESULTS: Three-hundred and seven patients underwent esophagectomy or esophagogastrectomy. Twenty-three (7.5%) developed esophageal anastomotic leaks. Eight of these patients (35%) died. Four of 23 (17%) patients had seemingly normal postoperative contrast studies. Factors potentially predictive of death included age (died, 72.8+/-8.3 years; survived, 65.3+/-8.8 years; p=0.063), location of anastomosis (cervical, 3/9 died; thoracic, 5/14 died; p=0.91), leak presentation (clinical, 6/12 died; contrast study, 2/11 died; p=0.11), time of leak (<7 days, 3/5 died; > or =7 days, 5/18 died; p=0.18), presence of gastric necrosis (necrosis, 3/3 died; no necrosis, 5/20 died; p=0.019), and treatment (surgical, 4/4 died; conservative, 4/19 died; p=0.005). CONCLUSIONS: Postoperative esophageal anastomotic leaks prove fatal in a significant number of cases. The lethal potential of cervical anastomotic leaks should not be underestimated. Gastric necrosis is an important predictor of subsequent death. Advanced age, early postoperative (<7 days) leakage, and clinically apparent signs of leakage may be predictive of death but these factors did not reach statistical significance in our study. Surgical treatment of esophageal anastomotic leaks is associated with subsequent death, but this relationship is unlikely to be causal; severely ill patients tend to be treated surgically.
Recent years have witnessed a meteoric increase in the adoption of smartphones. To manage information and features on such phones, Android provides a permission-based security model that requires each application to explicitly request permissions before it can be installed to run. In this paper, we analyze eight popular Android smartphones and discover that the stock phone images do not properly enforce the permission model. Several privileged permissions are unsafely exposed to other applications which do not need to request them for the actual use. To identify these leaked permissions or capabilities, we have developed a tool called Woodpecker. Our results with eight phone images show that among 13 privileged permissions examined so far, 11 were leaked, with individual phones leaking up to eight permissions. By exploiting them, an untrusted application can manage to wipe out the user data, send out SMS messages, or record user conversation on the affected phones – all without asking for any permission.
One aspect of security in mobile code is privacy: private (or secret) data should not be leaked to unauthorised agents. Most of the work on secure information flow has until recently only been concerned with detecting direct and indirect flows. Secret information can however be leaked to the attacker also through covert channels. It is very reasonable to assume that the attacker, even as an external observer, can monitor the timing (including termination) behaviour of the program. Thus to claim a program secure, the security analysis must take also these into account.
Permeation through nanometer pores is important in the design of materials for filtration and separation techniques and because of unusual fundamental behavior arising at the molecular scale. We found that submicrometer-thick membranes made from graphene oxide can be completely impermeable to liquids, vapors, and gases, including helium, but these membranes allow unimpeded permeation of water (H(2)O permeates through the membranes at least 10(10) times faster than He). We attribute these seemingly incompatible observations to a low-friction flow of a monolayer of water through two-dimensional capillaries formed by closely spaced graphene sheets. Diffusion of other molecules is blocked by reversible narrowing of the capillaries in low humidity and/or by their clogging with water.
暂无摘要（点击查看原文获取完整内容）
We discuss the idea that the accelerated universe could be the result of gravitational leakage into extra dimensions over Hubble distances rather than the consequence of a nonzero cosmological constant.
Methane emissions from U.S. and Canadian natural gas systems appear larger than official estimates.
OBJECTIVE: To examine the long-term oncological impact of anastomotic leakage (AL) after restorative surgery for colorectal cancer using meta-analytical methods. Outcomes evaluated were local recurrence, distant recurrence, and survival. BACKGROUND: Recurrence after potentially curative surgery for colorectal cancer remains a significant clinical problem and has a poor prognosis. AL may be a risk factor for disease recurrence, however available studies have been conflicting. A meta-analysis was conducted to investigate the impact of AL on disease recurrence and long-term survival. METHODS: Studies published between 1965 and 2009 evaluating the long-term oncological impact of AL were identified by an electronic literature search. Outcomes evaluated included local recurrence, distant recurrence, and cancer specific survival. Meta-analysis was performed using the DerSimonian-Laird random-effects model to compute odds ratio and 95% confidence intervals. Study heterogeneity was evaluated using Q statistics and I and publication bias assessed with funnel plots and Egger's test. RESULTS: Twenty-one studies comprising 13 prospective nonrandomized studies, 1 prospective randomized, and 7 retrospective studies met the inclusion criteria, yielding a total of 21,902 patients. For rectal anastomoses, the odd ratios (OR) of developing a local recurrence when there was AL was 2.05 (95% CI = 1.51-2.8; P = 0.0001). For studies describing both colon and rectal anastomoses, the OR of local recurrence when there was an AL was 2.9 (95% CI = 1.78-4.71; P < 0.001). The OR of developing a distant recurrence after AL was 1.38 (95% CI = 0.96-1.99; P = 0.083). Long term cancer specific mortality was significantly higher after AL with an OR of 1.75 (95% CI = 1.47-2.1; P = 0.0001). CONCLUSIONS: AL has a negative prognostic impact on local recurrence after restorative resection of rectal cancer. A significant association between colorectal AL and reduced long-term cancer specific survival was also noted. No association between AL and distant recurrence was found.
BACKGROUND: Anastomotic leak (AL) represents a dreaded complication following colorectal surgery, with a prevalence of 1-19 per cent. There remains a lack of consensus regarding factors that may predispose to AL and the relative risks associated with them. The objective was to perform a systematic review of the literature, focusing on the role of preoperative, intraoperative and postoperative factors in the development of colorectal ALs. METHODS: A systematic review was performed to identify adjustable and non-adjustable preoperative, intraoperative and postoperative factors in the pathogenesis of AL. Additionally, a severity grading system was proposed to guide treatment. RESULTS: Of 1707 papers screened, 451 fulfilled the criteria for inclusion in the review. Significant preoperative risk factors were: male sex, American Society of Anesthesiologists fitness grade above II, renal disease, co-morbidity and history of radiotherapy. Tumour-related factors were: distal site, size larger than 3 cm, advanced stage, emergency surgery and metastatic disease. Adjustable risk factors were: smoking, obesity, poor nutrition, alcohol excess, immunosuppressants and bevacizumab. Intraoperative risk factors were: blood loss/transfusion and duration of surgery more than 4 h. Stomas lessen the consequences but not the prevalence of AL. In the postoperative period, CT is the most commonly used imaging tool, with or without rectal contrast, and a C-reactive protein level exceeding 150 mg/l on day 3-5 is the most sensitive biochemical marker. A five-level classification system for AL severity and appropriate management is presented. CONCLUSION: Specific risk factors and their potential correction or indications for stoma were identified. An AL severity score is proposed to aid clinical decision-making.
Two-pore domain K(+) (K(2P)) channels give rise to leak (also called background) K(+) currents. The well-known role of background K(+) currents is to stabilize the negative resting membrane potential and counterbalance depolarization. However, it has become apparent in the past decade (during the detailed examination of the cloned and corresponding native K(2P) channel types) that this primary hyperpolarizing action is not performed passively. The K(2P) channels are regulated by a wide variety of voltage-independent factors. Basic physicochemical parameters (e.g., pH, temperature, membrane stretch) and also several intracellular signaling pathways substantially and specifically modulate the different members of the six K(2P) channel subfamilies (TWIK, TREK, TASK, TALK, THIK, and TRESK). The deep implication in diverse physiological processes, the circumscribed expression pattern of the different channels, and the interesting pharmacological profile brought the K(2P) channel family into the spotlight. In this review, we focus on the physiological roles of K(2P) channels in the most extensively investigated cell types, with special emphasis on the molecular mechanisms of channel regulation.
Tissue barriers that restrict passage of liquids, ions, and larger solutes are essential for the development of multicellular organisms. In simple organisms this allows distinct cell types to interface with the external environment. In more complex species, the diversity of cell types capable of forming barriers increases dramatically. Although the plasma membranes of these barrier-forming cells prevent flux of most hydrophilic solutes, the paracellular, or shunt, pathway between cells must also be sealed. This function is accomplished in vertebrates by the zonula occludens, or tight junction. The tight junction barrier is not absolute but is selectively permeable and is able to discriminate between solutes on the basis of size and charge. Many tight junction components have been identified over the past 20 years, and recent progress has provided new insights into the proteins and interactions that regulate structure and function. This review presents these data in a historical context and proposes an integrated model in which dynamic regulation of tight junction protein interactions determines barrier function.
Mitochondrial proton and electron leak have a major impact on mitochondrial coupling efficiency and production of reactive oxygen species. In the first part of this chapter, we address the molecular nature of the basal and inducible proton leak pathways, and their physiological importance. The basal leak is unregulated, and a major proportion can be attributed to mitochondrial anion carriers, whereas the proton leak through the lipid bilayer appears to be minor. The basal proton leak is cell-type specific and correlates with metabolic rate. The inducible leak through the ANT (adenine nucleotide translocase) and UCPs (uncoupling proteins) can be activated by fatty acids, superoxide or lipid peroxidation products. The physiological role of inducible leak through UCP1 in mammalian brown adipose tissue is heat production, whereas the roles of non-mammalian UCP1 and its paralogous proteins, in particular UCP2 and UCP3, are not yet resolved. The second part of the chapter focuses on the electron leak that occurs in the mitochondrial electron transport chain. Exit of electrons prior to the reduction of oxygen to water at cytochrome c oxidase causes superoxide production. As the mechanisms of electron leak are crucial to understanding their physiological relevance, we summarize the mechanisms and topology of electron leak from complexes I and III in studies using isolated mitochondria. We also highlight recent progress and challenges of assessing electron leak in the living cell. Finally, we emphasize the importance of proton and electron leak as therapeutic targets in body mass regulation and insulin secretion.
暂无摘要（点击查看原文获取完整内容）
BACKGROUND: Anastomotic leak after gastrointestinal surgery is an important postoperative event that leads to significant morbidity and mortality. Postoperative leak rates are frequently used as an indicator of the quality of surgical care provided. Comparison of rates between and within institutions depends on the use of standard definitions and methods of measurement of anastomotic leak. The aim of this study was to review the definition and measurement of anastomotic leak after oesophagogastric, hepatopancreaticobiliary and lower gastrointestinal surgery. METHODS: A systematic review was undertaken of the published literature. Searches were carried out on five bibliographical databases (Medline, Embase, The Cochrane Library, Cumulative Index for Nursing and Allied Health Literature and HealthSTAR) for English language articles published between 1993 and 1999. Articles were critically appraised by two independent reviewers and data on definition and measurement of anastomotic leak were extracted. RESULTS: Ninety-seven studies were reviewed and a total of 56 separate definitions of anastomotic leak were identified at three sites: upper gastrointestinal (13 definitions), hepatopancreaticobiliary (14) and lower gastrointestinal (29). The majority of studies used a combination of clinical features and radiological investigations to define and detect anastomotic leak. CONCLUSION: There is no universally accepted definition of anastomotic leak at any site. The definitions and values used to measure anastomotic failure vary extensively and preclude accurate comparison of rates between studies and institutions.
There have been no systematic empirical studies of the speed at which various kinds of technological information leak out to rival firms. To help fill this gap, data were obtained from 100 American firms. According to the results, information concerning development decisions is generally in the hands of rivals within about 12 to 18 months, on the average, and information concerning the detailed nature and operation of a new product or process generally leaks out within about a year. These results have important implications both for incentives for innovation and for public policies aimed at stemming the outflow of technology.
暂无摘要（点击查看原文获取完整内容）
暂无摘要（点击查看原文获取完整内容）
暂无摘要（点击查看原文获取完整内容）