88 research outputs found
Supporting complex workflows for data-intensive discovery reliably and efficiently
No full textScientific workflows have emerged as well-established pillars of large-scale computational science and appeared as torchbearers to formalize and structure a massive amount of complex heterogeneous data and accelerate scientific progress. Scientists of diverse domains can analyze their data by constructing scientific workflows as a useful paradigm to manage complex scientific computations. A workflow can analyze terabyte-scale datasets, contain numerous individual tasks, and coordinate between heterogeneous tasks with the help of scientific workflow management systems (SWfMSs). However, even for expert users, workflow creation is a complex task due to the dramatic growth of tools and data heterogeneity. Scientists are now more willing to publicly share scientific datasets and analysis pipelines in the interest of open science. As sharing of research data and resources increases in scientific communities, scientists can reuse existing workflows shared in several workflow repositories. Unfortunately, several challenges can prevent scientists from reusing those workflows, which hurts the purpose of the community-oriented knowledge base. In this thesis, we first identify the repositories that scientists use to share and reuse scientific workflows. Among several repositories, we find Galaxy repositories have numerous workflows, and Galaxy is the mostly used SWfMS. After selecting the Galaxy repositories, we attempt to explore the workflows and encounter several challenges in reusing them. We classify the reusability status (reusable/nonreusable). Based on the effort level, we further categorize the reusable workflows (reusable without modification, easily reusable, moderately difficult to reuse, and difficult to reuse). Upon failure, we record the associated challenges that prevent reusability. We also list the actions upon success. The challenges preventing reusability include tool upgrading, tool support unavailability, design flaws, incomplete workflows, failure to load a workflow, etc. We need to perform several actions to overcome the challenges. The actions include identifying proper input datasets, updating/upgrading tools, finding alternative tools support for obsolete tools, debugging to find the issue creating tools and connections and solving them, modifying tools connections, etc. Such challenges and our action list offer guidelines to future workflow composers to create better workflows with enhanced reusability. A SWfMS stores provenance data at different phases of a workflow life cycle, which can help workflow construction. This provenance data allows reproducibility and knowledge reuse in the scientific community. But, this provenance information is usually many times larger than the workflow and input data, and managing provenance data is growing in complexity with large-scale applications. In our second study, we document the challenges of provenance management and reuse in e-science, focusing primarily on scientific workflow approaches by exploring different SWfMSs and provenance management systems. We also investigate the ways to overcome the challenges. Creating a workflow is difficult but essential for data-intensive complex analysis, and the existing workflows have several challenges to be reused, so in our third study, we build a recommendation system to recommend tool(s) using machine learning approaches to help scientists create optimal, error-free, and efficient workflows by using existing reusable workflows in Galaxy workflow repositories. The findings from our studies and proposed techniques have the potential to simplify the data-intensive analysis, ensuring reliability and efficiency
A Fast and Scalable System to Visualize Contour Gradient from Spatio-temporal Data
No full textChanges in geological processes that span over the years may often go unnoticed due to their inherent noise and variability. Natural phenomena such as riverbank erosion, and climate change in general, is invisible to humans unless appropriate measures are taken to analyze the underlying data. Visualization helps geological sciences to generate scientific insights into such long-term geological events. Commonly used approaches such as side-by-side contour plots and spaghetti plots do not provide a clear idea about the historical spatial trends.
To overcome this challenge, we propose an image-gradient based approach called ContourDiff. ContourDiff overlays gradient vector over contour plots to analyze the trends of change across spatial regions and temporal domain. Our approach first aggregates for each location, its value differences from the neighboring points over the temporal domain, and then creates a vector field representing the prominent changes. Finally, it overlays the vectors (differential trends) along the contour paths, revealing the differential trends that the contour lines (isolines) experienced over time.
We designed an interface, where users can interact with the generated visualization to reveal changes and trends in geospatial data. We evaluated our system using real-life datasets, consisting of millions of data points, where the visualizations were generated in less than a minute in a single-threaded execution. We show the potential of the system in detecting subtle changes from almost identical images, describe implementation challenges, speed-up techniques, and scope for improvements. Our experimental results reveal that ContourDiff can reliably visualize the differential trends, and provide a new way to explore the change pattern in spatiotemporal data. The expert evaluation of our system using real-life WRF (Weather Research and Forecasting) model output reveals the potential of our technique to generate useful insights on the spatio-temporal trends of geospatial variables
TinySurveillance: A Low-power Event-based Surveillance Method for Unmanned Aerial Vehicles
No full textUnmanned Aerial Vehicles (UAVs) have always been faced with power management challenges. Managing power consumption becomes critical, especially in surveillance applications where the longer flight time results in wider coverage and a cheaper solution.
While most current studies focus on utilizing new models for improving event detection without considering the power constraints, our design's first priority is our platform's power efficiency. Implementing an algorithm on a portable device with minimal access to power supply sources requires special hardware and software considerations. An improved algorithm may need more powerful hardware, which can surge power consumption. Therefore, we aim to propose a method to be suitable for such devices with power consumption constraints.
In this work, we propose an event-driven surveillance method with an efficient video transmission algorithm that reduces power consumption while preserving image quality. The surveillance will start automatically once the low-power AI-based onboard processor detects the desired event. The drone repeatedly solves a classification problem by employing a lightweight deep learning algorithm. When the UAV detects the defined event, a sample image is sent to the server for validation. Afterwards, if the server validates the drone decision, the drone, which can be a UAV, starts sending a colored image accompanied by a group of N grayscale images. Then, in the server, the grayscale images will be colorized using a convolution neural network trained by the colored images. By adopting this method, the sent data rate decreases and the server's computation load increases. The former part results in a drop in the UAV's power consumption, which is our aim. In this work, an application of wildfire detection and surveillance has been implemented to show the proof of concept of the TinySurveillance method. Using four videos of similar scenarios with different spatial and temporal information that a UAV may face, with various spatial and temporal characteristics, we show the effectiveness of our method. Our results show that the power consumption of the onboard processing unit in detection mode will be reduced by at least 4 times, reaching a detection accuracy of 85%, while in surveillance mode, we can decrease the data transmission rate by almost 66% while achieving a competent image quality with PSNR_Avg of 41.35 dB, PSNR of 30.94 dB, and output frame rate of 5.2. Also, the reproduced images show the outstanding performance of the algorithm by generating colorized images identical to the original scenes. There are main features that affect our method's power consumption and output quality, like the number of grayscale images, sent video bitrate, learning rate, and video characteristics that are discussed comprehensively
Supporting complex workflows for data-intensive discovery reliably and efficiently
No full textScientific workflows have emerged as well-established pillars of large-scale computational science and appeared as torchbearers to formalize and structure a massive amount of complex heterogeneous data and accelerate scientific progress. Scientists of diverse domains can analyze their data by constructing scientific workflows as a useful paradigm to manage complex scientific computations. A workflow can analyze terabyte-scale datasets, contain numerous individual tasks, and coordinate between heterogeneous tasks with the help of scientific workflow management systems (SWfMSs). However, even for expert users, workflow creation is a complex task due to the dramatic growth of tools and data heterogeneity. Scientists are now more willing to publicly share scientific datasets and analysis pipelines in the interest of open science. As sharing of research data and resources increases in scientific communities, scientists can reuse existing workflows shared in several workflow repositories. Unfortunately, several challenges can prevent scientists from reusing those workflows, which hurts the purpose of the community-oriented knowledge base. In this thesis, we first identify the repositories that scientists use to share and reuse scientific workflows. Among several repositories, we find Galaxy repositories have numerous workflows, and Galaxy is the mostly used SWfMS. After selecting the Galaxy repositories, we attempt to explore the workflows and encounter several challenges in reusing them. We classify the reusability status (reusable/nonreusable). Based on the effort level, we further categorize the reusable workflows (reusable without modification, easily reusable, moderately difficult to reuse, and difficult to reuse). Upon failure, we record the associated challenges that prevent reusability. We also list the actions upon success. The challenges preventing reusability include tool upgrading, tool support unavailability, design flaws, incomplete workflows, failure to load a workflow, etc. We need to perform several actions to overcome the challenges. The actions include identifying proper input datasets, updating/upgrading tools, finding alternative tools support for obsolete tools, debugging to find the issue creating tools and connections and solving them, modifying tools connections, etc. Such challenges and our action list offer guidelines to future workflow composers to create better workflows with enhanced reusability. A SWfMS stores provenance data at different phases of a workflow life cycle, which can help workflow construction. This provenance data allows reproducibility and knowledge reuse in the scientific community. But, this provenance information is usually many times larger than the workflow and input data, and managing provenance data is growing in complexity with large-scale applications. In our second study, we document the challenges of provenance management and reuse in e-science, focusing primarily on scientific workflow approaches by exploring different SWfMSs and provenance management systems. We also investigate the ways to overcome the challenges. Creating a workflow is difficult but essential for data-intensive complex analysis, and the existing workflows have several challenges to be reused, so in our third study, we build a recommendation system to recommend tool(s) using machine learning approaches to help scientists create optimal, error-free, and efficient workflows by using existing reusable workflows in Galaxy workflow repositories. The findings from our studies and proposed techniques have the potential to simplify the data-intensive analysis, ensuring reliability and efficiency
Letter to the Editor and Author Response for "A systematic review and meta-analysis of randomized trials evaluating the efficacy of autologous skin cell suspensions for re-epithelialization of acute partial thickness burn injuries and split-thickness skin graft donor sites" by Bairagi, et al
No full textDear Editor,
We thank Holmes et al. for their interest in our systematic review and meta-analysis [1] and would like to submit the following response for their consideration:
1. Randomized trials eligible for evaluation in this study were limited to partial thickness burn injuries managed with an autologous skin cell suspension which was a pre-specified inclusion criteria for our systematic review, published on the PROSPERO International Prospective Register of Systematic Reviews prior to commencing screening of database results (PROSPERO Record ID = CRD42019133171) [2]. As such, findings from a study conducted in full-thickness burn wounds although significant [3], were excluded from evaluation as they were beyond the scope of our systematic review and meta-analysis. The role of autologous skin cell suspensions (ASCS) in the management of full thickness burn wounds, could be better understood in a future review examining full thickness wounds. The findings from the study conducted in full thickness burn wounds that was referred to by the authors of the letter to the editor could be included in such a review [3].No Full Tex
Improving the Spectral Efficiency of Modulation on Conjugate-Reciprocal Zeros (MOCZ) for Non-Coherent Short Packet Communications
No full textFuture internet of things (IoT) applications need to meet the stringent requirements of ultra high reliability and ultra low-latency. To meet the ultra low-latency requirements, the IoT networks will be employing the short data packets for data transmission between the devices. Employing the short data packet communications (SPCs) is not straightforward as there are several design problems related to the SPCs which still remain unsolved.
Since the block length for SPCs is finite; the channel estimation is a challenging problem. This is because the conventionally used known pilot symbols to estimate the channel will severely degrade the spectral efficiency of SPCs. Recently a novel non-coherent modulation technique named as modulation on conjugate reciprocal zeros (MOCZ) was proposed which supports the blind detection of transmitted data, i.e., detection without the knowledge of channel. It is also well known that SPCs suffers from data rate loss as compared to the channel capacity limit. Hence, in this thesis, we aim to increase the spectral efficiency of MOCZ.
We improve the spectral efficiency of MOCZ by proposing a technique named as spectrally efficient modulation on conjugate reciprocal zeros (SE-MOCZ) which combines MOCZ with a technique named as faster than Nyquist (FTN) Signaling. Hence; in SE-MOCZ, we end up transmitting the coefficients of MOCZ, modulated on T-orthogonal pulses, at a rate faster than the Nyquist limit, i.e., τT, instead of T, where 0 < τ < 1. That said, we intentionally introduce inter symbol interference (ISI) between the received samples of SE MOCZ. To partially remove the ISI, we introduce a discrete-time filter at the receiver. We further optimize the radius of complex zeros of SE-MOCZ in the presence of ISI. Simulation results show the gains of proposed SE-MOCZ in terms of spectral efficiency
Investigating the Quality Aspects of Crowd-Sourced Developer Forum: A Case Study of Stack Overflow
No full textTechnical question and answer (Q&A) websites have changed how developers seek information on the web and become more popular due to the shortcomings in official documentation and alternative knowledge sharing resources. Stack Overflow (SO) is one of the largest and most popular online Q&A websites for developers where they can share knowledge by answering questions and learn new skills by asking questions. Unfortunately, a large number of questions (up to 29%) are not answered at all, which might hurt the quality or purpose of this community-oriented knowledge base. In this thesis, we first attempt to detect the potentially unanswered questions during their submission using machine learning models. We compare unanswered and answered questions quantitatively and qualitatively. The quantitative analysis suggests that topics discussed in the question, the experience of the question submitter, and readability of question texts could often determine whether a question would be answered or not. Our qualitative study also reveals why the questions remain unanswered that could guide novice users to improve their questions. During analyzing the questions of SO, we see that many of them remain unanswered and unresolved because they contain such code segments that could potentially have programming issues (e.g., error, unexpected behavior); unfortunately, the issues could always not be reproduced by other users. This irreproducibility of issues might prevent questions of SO from getting answers or appropriate answers. In our second study, we thus conduct an exploratory study on the reproducibility of the issues discussed in questions and the correlation between issue reproducibility status (of questions) and corresponding answer meta-data such as the presence of an accepted answer. According to our analysis, a question with reproducible issues has at least three times higher chance of receiving an accepted answer than the question with irreproducible issues. However, users can improve the quality of questions and answers by editing. Unfortunately, such edits may be rejected (i.e., rollback) due to undesired modifications and ambiguities. We thus offer a comprehensive overview of reasons and ambiguities in the SO rollback edits. We identify 14 reasons for rollback edits and eight ambiguities that are often present in those edits. We also develop algorithms to detect ambiguities automatically. During the above studies, we find that about half of the questions that received working solutions have negative scores. About 18\% of the accepted answers also do not score the maximum votes. Furthermore, many users are complaining against the downvotes that are cast to their questions and answers. All these findings cast serious doubts on the reliability of the evaluation mechanism employed at SO. We thus concentrate on the assessment mechanism of SO to ensure a non-biased, reliable quality assessment mechanism of SO. This study compares the subjective assessment of questions with their objective assessment using 2.5 million questions and ten text analysis metrics. We also develop machine learning models to classify the promoted and discouraged questions and predict them during their submission time.
We believe that the findings from our studies and proposed techniques have the potential to (1) help the users to ask better questions with appropriate code examples, and (2) improve the editing and assessment mechanism of SO to promote better content quality
Code clone detection in obfuscated Android apps
No full textThe Android operating system has long become one of the main global smartphone operating systems. Both developers and malware authors often reuse code to expedite the process of creating new apps and malware samples. Code cloning is the most common way of reusing code in the process of developing Android apps. Finding code clones through the analysis of Android binary code is a challenging task that becomes more sophisticated when instances of code reuse are non-contiguous, reordered, or intertwined with other code. We introduce an approach for detecting cloned methods as well as small and non-contiguous code clones in obfuscated Android applications by simulating the execution of Android apps and then analyzing the subsequent execution traces. We first validate our approach’s ability on finding different types of code clones on 20 injected clones. Next we validate the resistance of our approach against obfuscation by comparing its results on a set of 1085 apps before and after code obfuscation. We obtain 78-87% similarity between the finding from non-obfuscated applications and four sets of obfuscated applications. We also investigated the presence of code clones among 1603 Android applications. We were able to find 44,776 code clones where 34% of code clones were seen from different applications and the rest are among different versions of an application. We also performed a comparative analysis between the clones found by our approach and the clones detected by Nicad on the source code of applications. Finally, we show a practical application of our approach for detecting variants of Android banking malware. Among 60,057 code clone clusters that are found among a dataset of banking malware, 92.9% of them were unique to one malware family or benign applications
Building on a Legacy : Working with users to revitalize the CRHM hydrological model
No full textA computer scientist's personal account of collaboratively migrating the CRHM hydrological modelling tool.Canada First Research Excellence FundNon-Peer ReviewedA computer scientist's personal account of the challenges involved in collaboratively migrating the CRHM hydrological modelling tool
- …
