Search

Prior studies have demonstrated the effectiveness of Deep Learning (DL) in automated software vulnerability detection. Graph Neural Networks (GNNs) have proven effective in learning the graph representations of source code and are commonly adopted by existing DL-based vulnerability detection methods. However, the existing methods are still limited by the fact that GNNs are essentially difficult to handle the connections between long-distance nodes in a code structure graph. Besides, they do not well exploit the multiple types of edges in a code structure graph (such as edges representing data flow and control flow). Consequently, despite achieving state-of-the-art performance, the existing GNN-based methods tend to fail to capture global information (i.e., long-range dependencies among nodes) of code graphs. To mitigate these issues, in this paper, we propose a novel vulnerability detection framework with grAph siMplification and enhanced graph rePresentation LEarning, named AMPLE. AMPLE mainly contains two parts: 1) graph simplification, which aims at reducing the distances between nodes by shrinking the node sizes of code structure graphs; 2) enhanced graph representation learning, which involves one edge-aware graph convolutional network module for fusing heterogeneous edge information into node representations and one kernel-scaled representation module for well capturing the relations between distant graph nodes. Experiments on three public benchmark datasets show that AMPLE outperforms the state-of-the-art methods by 0.39%-35.32% and 7.64%-199.81% with respect to the accuracy and F1 score metrics, respectively. The results demonstrate the effectiveness of AMPLE in learning global information of code graphs for vulnerability detection., Comment: 13 pages, 8 figures, Accepted for publication in the ICSE 23 Technical Track

IntroductionThe newly developed Composite Activity-related Risk of Falls Scale (CARFS) is designed to measure composite activity-related risk of falls (CARF) in relation to the activity-specific fear of falling and physical behavior. This study tested the reliability and validity of the CARFS in older people with various health statuses and persons with stroke or spinal cord injury.MethodsParticipants included 70 older adults, 38 persons with stroke, and 18 with spinal cord injury. They were first surveyed using a combined questionnaire including the CARFS and activity-specific balance confidence (ABC) scale in addition to items asking for personal and disease-related information, fall history, walking independence levels for examining internal consistency, ceiling and floor effects, and convergent validity in each participant group. One week after the initial survey, 33 older participants were reexamined using the CARFS to analyze test-retest reliability, where a minimal detectable change was found. Significance was set at α = 0.05 for all analyses.ResultsThe CARFS showed excellent test-retest reliability in the dimensions of fear of falling, physical behavior, and CARF [ICC (3,1) = 0.972, 0.994, and 0.994, respectively for their overall score], with a minimal detectable change of 3.944 in the older population. The internal consistency of CARFS items was excellent in the older participants, good in participants with stroke or spinal cord injury (Cronbach's alpha = 0.945, 0.843, 0.831 in each participant group, respectively). No ceiling and floor effects were demonstrated in the wide range of people. For the convergent validity, overall CARF score was significantly correlated with the average ABC score in each participant group (rho = −0.824, −0.761, and −0.601, respectively; p < 0.01), and was significantly correlated with walking independence levels in each participant group (rho = −0.636, −0.423, and −0.522, respectively; p < 0.01). It showed weak correlation with the number of previous falls only in participants with stroke (rho = 0.291, p = 0.076).ConclusionThe CARFS is a reliable and valid tool for measuring fall risk in older people and persons with stroke or spinal cord injury.

Unfair behaviors of Machine Learning (ML) software have garnered increasing attention and concern among software engineers. To tackle this issue, extensive research has been dedicated to conducting fairness testing of ML software, and this paper offers a comprehensive survey of existing studies in this field. We collect 100 papers and organize them based on the testing workflow (i.e., how to test) and testing components (i.e., what to test). Furthermore, we analyze the research focus, trends, and promising directions in the realm of fairness testing. We also identify widely-adopted datasets and open-source tools for fairness testing.

Software bias is an increasingly important operational concern for software engineers. We present a large-scale, comprehensive empirical study of 17 representative bias mitigation methods for Machine Learning (ML) classifiers, evaluated with 11 ML performance metrics (e.g., accuracy), 4 fairness metrics, and 20 types of fairness-performance trade-off assessment, applied to 8 widely-adopted software decision tasks. The empirical coverage is much more comprehensive, covering the largest numbers of bias mitigation methods, evaluation metrics, and fairness-performance trade-off measures compared to previous work on this important software property. We find that (1) the bias mitigation methods significantly decrease ML performance in 53% of the studied scenarios (ranging between 42%~66% according to different ML performance metrics); (2) the bias mitigation methods significantly improve fairness measured by the 4 used metrics in 46% of all the scenarios (ranging between 24%~59% according to different fairness metrics); (3) the bias mitigation methods even lead to decrease in both fairness and ML performance in 25% of the scenarios; (4) the effectiveness of the bias mitigation methods depends on tasks, models, the choice of protected attributes, and the set of metrics used to assess fairness and ML performance; (5) there is no bias mitigation method that can achieve the best trade-off in all the scenarios. The best method that we find outperforms other methods in 30% of the scenarios. Researchers and practitioners need to choose the bias mitigation method best suited to their intended application scenario(s)., Comment: Accepted by ACM Transactions on Software Engineering and Methodology (TOSEM 2023). Please include TOSEM in any citations

Additional file 1.

With little to no parallel data available for programming languages, unsupervised methods are well-suited to source code translation. However, the majority of unsupervised machine translation approaches rely on back-translation, a method developed in the context of natural language translation and one that inherently involves training on noisy inputs. Unfortunately, source code is highly sensitive to small changes; a single token can result in compilation failures or erroneous programs, unlike natural languages where small inaccuracies may not change the meaning of a sentence. To address this issue, we propose to leverage an automated unit-testing system to filter out invalid translations, thereby creating a fully tested parallel corpus. We found that fine-tuning an unsupervised model with this filtered data set significantly reduces the noise in the translations so-generated, comfortably outperforming the state-of-the-art for all language pairs studied. In particular, for Java $\to$ Python and Python $\to$ C++ we outperform the best previous methods by more than 16% and 24% respectively, reducing the error rate by more than 35%.

Complementarity of variable renewable energy sources at multiple temporal scales is important in order to ensure reliability of a decarbonizing energy system. In this study, we investigate the hypothesis that a decrease in the fraction of precipitation falling as snow (SWE/P) would increase monthly complementarity of hydro and solar power generation in the western U.S. With a focus on 123 dams responsible for 93% of generation, we found that these resources are seasonally complementary at about half of dams, as indicated by the sign of correlation coefficients (ρ). As hypothesized, average SWE/P at individual dams was generally positively correlated with ρ, but the dependence of ρ on SWE/P was non-linear and SWE/P only explained a modest portion of the variance in complementarity. At each dam, the dependence of annual ρ on interannual variations in SWE/P between 2002–2020 was assessed; these relationships were positive at 72% of dams but not statistically significant at the level of individual dams. Finally, at the system scale ρ was significantly related to SWE/P, with a stronger relationship observed than the dependence of total hydropower generation on SWE/P. Notably, the system-scale relationship between ρ and SWE/P changed dramatically in the latter part of the temporal domain (2012–2020), with a much steeper slope and greater fraction of variance explained by SWE/P. These results illustrate the historical relationship between SWE/P, monthly complementarity of hydro and solar power, complexities of these relationships due to snow and watershed hydrology and reservoir management, and a change in the observed relationship between SWE/P and hydropower generation timing. To the extent that hydro and solar power generation complementarity is responsive to SWE/P, expected declines in SWE/P may indicate greater seasonal complementarity but reduced hydropower available for load-balancing when solar power generation is highest.

There are few data on outcomes for patients with metastatic urothelial carcinoma (MUC) who receive chemotherapy (CT) after progression on immune checkpoint inhibitors (ICIs). We carried out a retrospective single-centre analysis of MUC patients who progressed after ICI and then received CT. Patients fell into two groups: CT-naive (no prior-CT) and CT-pretreated (platinum-based CT followed by ICI on progression). The response rate (RR), progression-free survival (PFS), and duration of response (DOR) were assessed. A total of 29 patients received CT following progression on ICI. The median follow-up was 17.0mo (interquartile range 9.1-20.5mo). In the CT-naive group (n=17), 53% had a partial response, 18% had stable disease, and 29% had progressive disease. In the CT-pretreated group (n=12) 17% had a partial response, 67% had stable disease, and 16% had progressive disease. The median PFS was 6.4mo (95% confidence interval [CI] 3.8-9.1) in the CT-naive and 4.4mo (95% CI 1.5-7.3) in the CT-pretreated group. The median DOR was 8.1mo (range 5.1-11.1) among the ten patients with a response to CT after ICI in both groups. Some 38% of patients in the CT-naive and 17% in the CT-pretreated group had dose reductions on post-ICI CT. CT and ICI can be sequenced after previous chemotherapy exposure, although this does not induce long-term durable remissions in most patients. PATIENT SUMMARY: We looked at outcomes for patients with metastatic bladder cancer who received chemotherapy after the cancer got worse while on immunotherapy. We found that patients can be safely treated with further chemotherapy. However, the positive effects of chemotherapy will not be durable in the majority of patients.

Analysis of the ecosystem service value (ESV) associated with the observed land use changes is critically important for environmental assessment and sustainable urban development. This research evaluate the effects of land use on ESV in Chengdu based on remotely sensed data and GIS, assessment model, transfer and flows model, sensitivity and variable ratio function. The results show that the ESV is mainly composed of forest land and cultivated land. ESV substantially and continuously decreased by 92 × 108 Chinese Yuan (CNY) or about 45 % in Chengdu during the study period and ESV from urban areas in the city center and its surroundings were much lower than those from the natural landscape in the northern and western Longmenshan mountainous region. Land use changes were the major causes of the observed transfer and flows in ESV. The flows of ESV both positive and negative were associated with observed land use changes, but overall land use changes had a negative impact on ESV. The results further indicate that the coefficient change relative to the ESV was inelastic, and the relative sensitivity index was as follows: forest land > cultivated land > water areas, but changes in water area had the largest influence and amplification effect on the ESV. This research offers important insights for use in planning fast urbanizing regions and can assist to achieve more sustainable land use and policy making.

The practical and symbolic significance of the sun and the moon was coded for a world sample of societies. Methods included use of the eHRAF World Cultures (hereafter eHRAF) Word Search function and the coding of relevant eHRAF materials. The sun, objectively of greater importance than the moon for many aspects of human existence, was expected to be ethnographically more prominent than the moon, and this was found in 55 of 60 cases. There was also a strong tendency for ethnographic mentions of the moon to be made in connection with magico-religious ideas, whereas statements about the sun tended to be more equally divided between the magico-religious and techno-econo-environmental domains. In exploratory investigation, it was found that both sun and moon were frequently mentioned in the ethnographies of foraging societies, which in turn were found to be at the low end of a scale of cultural complexity. Discussion centered on interpretation of the findings and on suggestions for further research.

Sudden cardiac death (SCD) remains an unsolved problem in the twenty-first century. It is often due to rapid onset, ventricular arrhythmias caused by a number of different clinical conditions. A proportion of SCD patients have identifiable diseases such as cardiomyopathies, but for others, the causes are unknown. Viral myocarditis is becoming increasingly recognized as a contributor to unexplained mortality, and is thought to be a major cause of SCD in the first two decades of life. Myocardial inflammation, ion channel dysfunction, electrophysiological, and structural remodeling may play important roles in generating life-threatening arrhythmias. The aim of this review article is to examine the electrophysiology of action potential conduction and repolarization and the mechanisms by which their derangements lead to triggered and reentrant arrhythmogenesis. By synthesizing experimental evidence from pre-clinical and clinical studies, a framework of how host (inflammation), and viral (altered cellular signaling) factors can induce ion electrophysiological and structural remodeling is illustrated. Current pharmacological options are mainly supportive, which may be accompanied by mechanical circulatory support. Heart transplantation is the only curative option in the worst case scenario. Future strategies for the management of viral myocarditis are discussed.

The Sec-dependent translocation pathway that involves the essential SecA protein and the membrane-bound SecYEG translocon is used to export many proteins across the cytoplasmic membrane. Recently, several pathogenic bacteria, including Mycobacterium tuberculosis , were shown to possess two SecA homologs, SecA1 and SecA2. SecA1 is essential for general protein export. SecA2 is specific for a subset of exported proteins and is important for M. tuberculosis virulence. The enzymatic activities of two SecA proteins from the same microorganism have not been defined for any bacteria. Here, M. tuberculosis SecA1 and SecA2 are shown to bind ATP with high affinity, though the affinity of SecA1 for ATP is weaker than that of SecA2 or Escherichia coli SecA. Amino acid substitution of arginine or alanine for the conserved lysine in the Walker A motif of SecA2 eliminated ATP binding. We used the SecA2(K115R) variant to show that ATP binding was necessary for the SecA2 function of promoting intracellular growth of M. tuberculosis in macrophages. These results are the first to show the importance of ATPase activity in the function of accessory SecA2 proteins.

Pb(Zr 0.52 Ti 0.48 )O 3 (PZT) thin films were prepared on Pt/Ti/SiO 2 /Si substrates by hybrid processing: sol-gel method and pulsed laser deposition. The temperature of postdeposition annealing in hybrid processing is 650°C, and is lower than that in the case of direct film deposition by pulsed laser deposition on a Pt/Ti/SiO 2 /Si substrate. The preferred orientation of the PZT films obtained by hybrid processing can be controlled using the seed layer obtained by the sol-gel process. The TEM image showed that the PZT films have a polycrystalline columnar microstructure extending throughout the thickness of the film and no sharp interface was observed between the layers obtained by the sol-gel method and the pulsed laser deposition process. Electrical properties of the films were evaluated by measuring their P-E hysteresis loops and dielectric constants. The 1-μm-thick PZT films fabricated by hybrid processing consist of mainly the perovskite phase with a (111)-preferred orientation and have good ferroelectric properties. The ferroelectric parameters were remanent polarization P r = 23.6 μC/cm 2 , and coercive field E c = 54.8 kV/cm.

Seven SUNCT patients (six men, one woman) took part in this study. In four patients, respiratory variables were compared during and outside attacks. In five patients, peripheral chemosensitivity was tested and compared with a control group matched with respect to age, sex, and smoking habits. The results indicate that SUNCT patients hyperventilate during attacks. Moreover, they appear to hyperventilate slightly under basal conditions. The tests for peripheral chemoreceptor activity indicated no differences between the SUNCT and the control groups except for one variable, namely the mean ventilatory response to a single breath of 13% CO2. It is possible that this indicates a blunted response of the peripheral chemoreceptors. On the other hand, it may also represent a chance finding, since none of the other results presented suggested such a conclusion, and the size of the test group was very small. The results do not indicate that a reduction in oxygen saturation can trigger SUNCT since low levels of oxygen saturation were only rarely accompanied by SUNCT, whereas many attacks were not associated with any appreciable lowering in arterial oxygen saturation.