Research Paper] Static JavaScript Call Graphs: A Comparative Study

Static call graphs are a fundamental building block of program analysis. However, differences in call-graph construction and the use of specific language features can yield unsoundness and imprecision. Call-graph analyses are evaluated using measures of precision and recall, but this is hard when a ground truth for real-world programs is generally unobtainable. In this work, we propose to use carefully constructed dynamic baselines based on fixed entry points and input corpora. The creation of this dynamic baseline is posed as an approximation of the ground truth---an optimization problem. We use manual extension and coverage-guided fuzzing for creating suitable input corpora. With these dynamic baselines, we study call-graph quality of multiple algorithms and implementations using four real-world Java programs. We find that our methodology provides valuable insights into call-graph quality and how to measure it. With this work, we provide a novel methodology to advance the field of static program analysis as we assess the computation of one of its core data structures---the call graph.

The use of pointers presents serious problems for software productivity tools for software understanding, restructuring, and testing. Pointers enable indirect memory accesses through pointer dereferences, as well as indirect procedure calls (e.g., through function pointers in C). Such indirect accesses and calls can be disambiguated with pointer analysis. In this paper we evaluate the precision of one specific pointer analysis (the FA pointer analysis by Zhang et al.) for the purposes of call graph construction for C programs with function pointers. The analysis is incorporated in a production-strength code-browsing tool from Siemens Corporate Research in which the program call graph is used as a primary tool for code understanding. The FA pointer analysis uses an inexpensive, almost-linear, flow- and context-insensitive algorithm. To measure analysis precision, we compare the call graph constructed by this analysis with the most precise call graph obtainable by a large category of existing pointer analyses. Surprisingly, for all our data programs the FA analysis achieves the best possible precision. This result indicates that for the purposes of call graph construction, inexpensive pointer analyses may provide precision comparable to the precision of expensive pointer analyses.

Language models for coding are shifting their focus from function-level to repository-level, with complex function invocations. We introduce CG-Bench, the first manually constructed benchmark that measures the ability to understand call graphs for language models. This benchmark contains 104 call sites and related code snippets associated with call chains from 7 representative open-source C/C++ projects. Language models are tasked with inference the calling targets from them. We evaluated four popular language models on CG-Bench. Surprisingly, all four models with different prompt settings achieve accuracy greater than 50% and Deepseek-6.7b with few-shot prompts reaches 69.70%. We further show four findings from a micro study, which demonstrates that using language models for call graph construction is promising and the performance can be improved by prompt hacking, removing irrelevant information, etc.

Call graph construction serves as a foundational component in Python static analysis and software engineering. However, it faces significant challenges due to the language's unique characteristics. That is, the dynamic typing system, extensive use of third-party libraries, and runtime-dependent behaviors in Python create substantial obstacles for traditional static analysis methods. Although these methods typically achieve high precision, they suffer from low recall, making it difficult to strike a balance between soundness and precision. In this paper, we present InferCG , a hybrid framework that addresses these limitations by combining coarse-grained static analysis with large language model (LLM)–based semantic reasoning. InferCG first constructs a high-recall candidate call set using permissive name- and arity-based matching, and then employs powerful LLMs such as DeepSeek and Qwen to semantically filter implausible edges. This two-stage design balances recall and precision without requiring dynamic execution or extensive type annotations. Experiments on curated open-source Python projects and the DyPyBench benchmark demonstrate that InferCG outperforms state-of-the-art static tools, which achieves up to 13.9% improvement in average recall and a 5.0% gain in average F1 score over PyCG. Moreover, in a real-world vulnerable-function detection scenario, InferCG enables more accurate reachability analysis than dependency-based tools.

Call graphs generated by profiling tools are critical to dissecting the performance of parallel programs. Although many mature and sophisticated profiling tools record call graph data, each tool is different in its runtime overheads, memory consumption, and output data generated. In this work, we perform a comparative evaluation study on the call graph data generation capabilities of several popular profiling tools – Caliper, HPCToolkit, TAU, and Score-P. We evaluate their runtime overheads, memory consumption, and generated call graph data (size and quality). We perform this comparison empirically by executing several proxy applications, AMG, LULESH, and Quicksilver on a parallel cluster. Our results show which tool results in the lowest overheads and produces the most meaningful call graph data under different conditions.KeywordsProfiling toolsCall graphPerformance analysisParallel performanceMeasurement

Purpose – This paper aims to find an objects representation scheme with high precision and to compute the objects’ separation distance effectively in final analysis. Proximity queries have been used widely in robot trajectory planning, automatic assembly planning, virtual surgery and many other applications. The core of proximity query is the precise computation of (minimum) separation distance in narrow phase, and specific object representation scheme corresponds to different methods of separation distance computation. Design/methodology/approach – In this paper, a second-order cone programming (SOCP)-based (minimum) separation distance computation algorithm was proposed. It treats convex superquadrics, descriptive primitives of complex object as the study objects. The separation distance between two convex superquadrics was written as a general nonlinear programming (NLP) problem with superquadric constraints and then transformed into an SOCP problem with the method of conic formulation of superquadric constraints. Finally, a primal-dual interior point method embedded in MOSEK was used for solving the SOCP problem. Findings – The proposed algorithm achieved exact separation distance computation between convex superquadrics, with a relative error of 10-6. It is particularly suitable for proximity queries in narrow phase of static collision detection algorithms. Further, the proposed algorithm achieved continuous collision detection between rectilinear translation superquadrics. Originality/value – The proposed algorithm in narrow phase of static collision detection algorithms makes objects’ separation distance effectively computed. Proximity queries are easy and more accurate to perform in this way.

Since the start of Roll Back Malaria (RBM) in 1998 funding for malaria control has increased dramatically, resulting in the current peak of $2.5billion spent on global malaria control annually. Vector control has been a major source of expenditure, with the focus in sub-Saharan Africa being free Long-Lasting Insecticidal Net (LLIN) distribution and Indoor Residual Spraying (IRS). Use of pyrethroid insecticides in agriculture and rapid scaling up of pyrethroid LLINs and IRS for malaria vector control has led to the development and spread of pyrethroid resistance in Anopheles gambiae malaria vectors. In community use, the level of insecticide resistance at which malaria control is compromised remains uncertain, but experimental hut trials in Benin, an area of high frequency pyrethroid resistance, showed that holed pyrethroid Insecticide Treated Nets (ITNs) failed to protect sleepers from being bitten and no longer had a mass killing effect on malaria vectors. If LLINs and IRS are to remain effective it is essential that new public health insecticides are developed to address the growing problem of resistance. All insecticides that are currently recommended by the World Health Organization Pesticide Evaluation Scheme (WHOPES) for LLIN or IRS belong to just four classes of chemistry that act on nerve and muscle targets; namely pyrethroid, organophosphate (OP), carbamate, and organochlorine (DDT). The Global Plan for Insecticide Resistance Management (GPIRM) states that in areas of pyrethroid resistance or high LLIN coverage, alternative insecticide classes should be used for IRS in a rotation. Rotation of insecticides is very difficult to implement due to a lack of new public health insecticides. The Stockholm Convention on Persistent Organic Pollutants (POPs) came into effect in 2004, yet the use of DDT (classified as a POP) for malaria control has been allowed to continue under exemption since then due to a perceived absence of equally effective and efficient alternatives. Alternative classes of insecticide for IRS such as pirimiphos-methyl (OP) and bendiocarb (carbamate) have a relatively short residual duration of action (2-6 months according to WHOPES). In areas of year-round transmission, multiple spray cycles are required resulting in significantly higher costs for malaria control programs and user fatigue. For continued cost-effectiveness of IRS programs it is important to develop new longer-lasting formulations of currently available insecticides, while also developing insecticides with new modes of action. Pyrethroids are the only insecticides that are currently recommended by WHOPES for LLIN. Therefore, it is essential to develop and evaluate new insecticides for LLIN before effectiveness of pyrethroid LLIN is compromised.
\n6
\nThis thesis consisted of a sequence of tests to evaluate the efficacy of several new formulations of WHOPES recommended insecticides and novel insecticides both in the laboratory and against wild mosquitoes entering experimental huts.
\nSpecifically these studies have shown that:
\n Addition of eave baffles in experimental huts succeeded in reducing the potential for mosquito escape and is preferable to the assumption of doubling veranda catch to allow for unrecorded escapes (research paper 2).
\n A Capsule Suspension (CS) formulation of pirmiphos-methyl used for IRS showed a significant improvement in terms of longevity on mud, concrete and plywood when compared with the previously recommended Emulsifiable Concentrate (EC) formulation in laboratory and experimental hut bioassays (research paper 3).
\n A new formulation of deltamethrin with polymeric binder (SC-PE) for IRS showed only a slight improvement over the existing Water Dispersible Granules (WG) formulation in bioassays, but both formulations equalled DDT in experimental huts and should provide annual mosquito control. Deltamethrin SC-PE or WG should only be considered for use by malaria control programs where there is low pyrethroid LLIN coverage (research paper 4).
\n In experimental hut trials, chlorfenapyr (pyrrole) IRS was equivalent to alphacypermethrin against pyrethroid susceptible An. arabiensis but superior against pyrethroid-resistant Cx. quinquefasciatus. The unique non-neurological mode of action shows no cross-resistance to existing resistance mechanisms and should be successful for control of pyrethroid resistant mosquitoes (research paper 5).
\n In experimental hut trials, chlorfenapyr ITNs produced relatively high mortality rates of pyrethroid susceptible An. arabiensis but due to low irritability there was only a small reduction in blood-feeding (research paper 8). Mortality rates were similar to those produced by deltamethrin ITN.
\n Unlike neurotoxic insecticides, such as pyrethroids and carbamates, chlorfenapyr owes its toxicity to the disruption of molecular pathways which enable cellular respiration to occur. Conventional 3 minute contact bioassay based on WHOPES guidelines is suitable for pyrethroids but does not predict field performance of
\n7
\nchlorfenapyr, which is metabolic in nature and sensitive to temperature and the phase of the insect’s circadian activity rhythm (research paper 9).
\n Combining chlorfenapyr with a more excito-repellent pyrethroid on mosquito nets produced higher levels of blood-feeding inhibition than chlorfenapyr alone, in tunnel tests with both pyrethroid susceptible and resistant strains of Cx. quinquefasciatus (research paper 10).
\n Restricting insecticide to particular surfaces of the nets (top only or sides only) indicated that An. arabiensis contacts both the top and sides of a mosquito net during host-seeking behaviour. These results support the rationale behind the ‘2-in-1’ mosquito net, in which the top of the net is treated with a non-pyrethroid insecticide and the sides with pyrethroid (research paper 11).

Language analysis of scientific documents and analysis of source code have been done independently in the past. This work presents a network architecture and a self-supervised training approach to find alignments between published computer science research papers and their corresponding public source code by learning a representation of encodings from transformers, from which source code can be enriched with helpful information. We present our ideas, findings and plans for upcoming research.

This thesis investigates gender inequalities in healthy ageing among the older English population, using data from the English Longitudinal Study of Ageing. The research aims were achieved by completing the following steps: (i) healthy ageing was intended as advancing to the later stages of the life course without disability; where disability was first theoretically conceptualized and then measured using severity levels that were identified empirically; (ii) gender inequalities in healthy ageing were assessed by studying whether the association between disability and mortality observed over the course of a decade differed between men and women; and (iii) disability and mortality were combined into a summary measure of population health -disability-free life expectancy- in order to estimate how expectancies of healthy life have changed over a decade across the two genders. The work is structured in four papers, denoted Research Paper I-IV. Research Paper I, a systematic literature review of studies analysing inequalities in health expectancy among the older population, inspired the direction taken by this thesis, as it identified gaps and open questions to be addressed to aid the understanding of the dynamics of healthy ageing. Research Paper II attempted to develop an approach to answer some of these questions. First, a solid and theoretically grounded definition of disability was proposed, based on the WHO's International Classification of Functioning Disability and Health (ICF), and in contrast to the data-dependent (and therefore heterogeneous) measures used in the literature. Then, using this definition, explanations of the gender paradox in health and mortality were attempted by analysing whether the association of disability with mortality differed between women and men over the period for which data were available (2002-2012). In Research Paper III the definition of disability elaborated in Research Paper II was used to foster and advance the debate on the usefulness and relevance of adopting a finer categorization of disability, and discuss why it is important to go beyond a binary classification, and to identify the appropriate number of disability levels that is most useful for research purposes. Based on these conclusions, the final aim of this thesis was accomplished in Research Paper IV, which studied the trends in disability-free life expectancy in England over the last decade, comparing the changes experienced by men and women at each severity level of disability. The collective findings of this thesis highlight the importance of defining disability in a consistent and comprehensive way as well as considering different severity levels. This work provides robust empirical evidence for theories of population health change over a decade in the English setting, with gender differences in healthy ageing, and directions of population health changes, found to vary across disability levels.

Assessment of research papers in medical university staff evaluation

Node localization, which is formulated as an unconstrained NP-hard optimization problem, is considered as one of the most significant issues of wireless sensor networks (WSNs). Recently, many swarm intelligent algorithms (SIAs) were applied to solve this problem. This study aimed to determine node location with high precision by SIA and presented a new localization algorithm named LMQPDV-hop. In LMQPDV-hop, an improved DV-Hop was employed as an underground mechanism to gather the estimation distance, in which the average hop distance was modified by a defined weight to reduce the distance errors among nodes. Furthermore, an efficient quantum-behaved particle swarm optimization algorithm (QPSO), named LMQPSO, was developed to find the best coordinates of unknown nodes. In LMQPSO, the memetic algorithm (MA) and Lévy flight were introduced into QPSO to enhance the global searching ability and a new fast local search rule was designed to speed up the convergence. Extensive simulations were conducted on different WSN deployment scenarios to evaluate the performance of the new algorithm and the results show that the new algorithm can effectively improve position precision.

The quick spurt in online transactions has also brought with it a parallel surge in fraud. With digital payments, the scope for fraud remains very high. Credit card fraud, among others, can cause heavy losses to customers and erode the confidence of consumers in Internet transactions. Furthermore, the detection of fraudsters in the online world poses big challenges. For one thing, there is an imbalance in data: fraud transactions are very few compared to genuine transactions. In this research paper, we intend to classify fraud through a supervised machine learning method. SVM is utilized for classification purposes. Based on the available dataset, we performed data analysis to obtain valuable information concerning fraud detection. To solve the problem of data imbalance, we first preprocessed the raw data by randomly choosing some legitimate transactions and normalizing the features. We also used feature selection and scaling methods to improve the accuracy of the model. After training the SVM on the sanitized dataset, we tested the model using performance measures like accuracy, precision, recall, and F1-score. These measures are especially important when working with skewed data. The findings show that the SVM model can detect fraudulent transactions with high precision and a good rate of recall. This shows that it can assist in reducing false alarms while being able to detect most fraud cases. In addition, we touch upon the tradeoff between false negatives and false positives because both pose very significant implications within financial institutions.

Easy Home or Home automation plays a very important role in modern era because of its flexibility in using it at different places with high precision which will save money and time by decreasing human hard work. Prime focus of this technology is to control the household equipment’s like light, fan, door, AC etc. automatically. This research paper has detailed information on Home Automation and Security System using Arduino, GSM and how we can control home appliances using Android application. Whenever a person will enter into the house then the count of the number of persons entering in the house will be incremented, in Home Automation mode applicances will be turned on whereas in security light will be turned on along with the alarm. The count of the number of persons entering the house is also displayed on the LCD screen. In Home Automation mode when the room will become empty i.e. the count of persons reduces to zero then the applicances will be turned off making the system power efficient. Moreover a person can control his home appliances by using an android application present in his mobile phone which will reduce the human hard work. At the same time if anyone enters while security mode is on a SMS will be sent to house owner’s mobile phone which will indicate the presence of a person inside the house.The alarm can be turned of using SMS or Android application.

The applications of Artificial Intelligence ie AI show diversity in any fields. On the other hand, research of the predicting heat transfer regardless of single-phase or two-phase flow is still untouched. Therefore, we have confirmed usefulness using AI’s deep learning function on horizontal flow boiling heat transfer in flowing mini-channel that is actively researched. The effect of the surface tension in the mini-channel is large compared with conventional large tubes, and then the heat transfer mechanism is very complicated. For this reason, the numerical correlations of many existing researchers the prediction result is not good. However, the mechanistic correlation based on the visualization experiment, which the authors' research group published several years ago has very high precision. Therefore, in this research paper, we confirmed the effectiveness of using deep learning for predicting of the boiling heat transfer in mini-channel while comparing our correlation.

Fault trees constitute one of the essential formalisms for static safety analysis of various industrial systems. Dynamic fault trees (DFT) enrich the formalism by support for time-dependent behaviour, e.g., repairs or dynamic dependencies. This enables more realistic and more precise modelling, and can thereby avoid overly pessimistic analysis results. But analysis of DFT is so far limited to substantially smaller models than those required for instance in the domain of nuclear power safety. This paper considers so called SD fault trees, where the user is free to express each equipment failure either statically, without modelling temporal information, or dynamically, allowing repairs and other timed interdependencies. We introduce an analysis algorithm for an important subclass of SD fault trees. The algorithm employs automatic abstraction techniques effectively, and thereby scales similarly to static analysis algorithms, albeit allowing for a more realistic modelling and analysis. We demonstrate the applicability of the method by an experimental evaluation on fault trees of nuclear power plants.