arXiv subjects

Tristan Glatard

Publications and source records attributed to Tristan Glatard.

At least 37 records · Page 2Linked to original sources

arXiv2022

PyTracer: Automatically profiling numerical instabilities in Python

Numerical stability is a crucial requirement of reliable scientific computing. However, despite the pervasiveness of Python in data science, analyzing large Python programs remains challenging due to the lack of scalable numerical analysis tools available for this language. To fill this gap, we developed PyTracer, a profiler to quantify numerical instability in Python applications. PyTracer transparently instruments Python code to produce numerical traces and visualize them interactively in a Plotly dashboard. We designed PyTracer to be agnostic to numerical noise model, allowing for tool evaluation through Monte-Carlo Arithmetic, random rounding, random data perturbation, or structured noise for a particular application. We illustrate PyTracer's capabilities by testing the numerical stability of key functions in both SciPy and Scikit-learn, two dominant Python libraries for mathematical modeling. Through these evaluations, we demonstrate PyTracer as a scalable, automatic, and generic framework for numerical profiling in Python.

cs.MS↗

arXiv2021

Data Augmentation Through Monte Carlo Arithmetic Leads to More Generalizable Classification in Connectomics

Machine learning models are commonly applied to human brain imaging datasets in an effort to associate function or structure with behaviour, health, or other individual phenotypes. Such models often rely on low-dimensional maps generated by complex processing pipelines. However, the numerical instabilities inherent to pipelines limit the fidelity of these maps and introduce computational bias. Monte Carlo Arithmetic, a technique for introducing controlled amounts of numerical noise, was used to perturb a structural connectome estimation pipeline, ultimately producing a range of plausible networks for each sample. The variability in the perturbed networks was captured in an augmented dataset, which was then used for an age classification task. We found that resampling brain networks across a series of such numerically perturbed outcomes led to improved performance in all tested classifiers, preprocessing strategies, and dimensionality reduction techniques. Importantly, we find that this benefit does not hinge on a large number of perturbations, suggesting that even minimally perturbing a dataset adds meaningful variance which can be captured in the subsequently designed models.

q-bio.QM↗

arXiv2021

The benefits of prefetching for large-scale cloud-based neuroimaging analysis workflows

To support the growing demands of neuroscience applications, researchers are transitioning to cloud computing for its scalable, robust and elastic infrastructure. Nevertheless, large datasets residing in object stores may result in significant data transfer overheads during workflow execution. Prefetching, a method to mitigate the cost of reading in mixed workloads, masks data transfer costs within processing time of prior tasks. We present an implementation of "Rolling Prefetch", a Python library that implements a particular form of prefetching from AWS S3 object store, and we quantify its benefits. Rolling Prefetch extends S3Fs, a Python library exposing AWS S3 functionality via a file object, to add prefetch capabilities. In measured analysis performance of a 500 GB brain connectivity dataset stored on S3, we found that prefetching provides significant speed-ups of up to 1.86x, even in applications consisting entirely of data loading. The observed speed-up values are consistent with our theoretical analysis. Our results demonstrate the usefulness of prefetching for scientific data processing on cloud infrastructures and provide an implementation applicable to various application domains.

cs.DC↗

arXiv2021

A Recommender System for Scientific Datasets and Analysis Pipelines

Scientific datasets and analysis pipelines are increasingly being shared publicly in the interest of open science. However, mechanisms are lacking to reliably identify which pipelines and datasets can appropriately be used together. Given the increasing number of high-quality public datasets and pipelines, this lack of clear compatibility threatens the findability and reusability of these resources. We investigate the feasibility of a collaborative filtering system to recommend pipelines and datasets based on provenance records from previous executions. We evaluate our system using datasets and pipelines extracted from the Canadian Open Neuroscience Platform, a national initiative for open neuroscience. The recommendations provided by our system (AUC$=0.83$) are significantly better than chance and outperform recommendations made by domain experts using their previous knowledge as well as pipeline and dataset descriptions (AUC$=0.63$). In particular, domain experts often neglect low-level technical aspects of a pipeline-dataset interaction, such as the level of pre-processing, which are captured by a provenance-based system. We conclude that provenance-based pipeline and dataset recommenders are feasible and beneficial to the sharing and usage of open-science resources. Future work will focus on the collection of more comprehensive provenance traces, and on deploying the system in production.

cs.IR↗

arXiv2021

Accurate simulation of operating system updates in neuroimaging using Monte-Carlo arithmetic

Operating system (OS) updates introduce numerical perturbations that impact the reproducibility of computational pipelines. In neuroimaging, this has important practical implications on the validity of computational results, particularly when obtained in systems such as high-performance computing clusters where the experimenter does not control software updates. We present a framework to reproduce the variability induced by OS updates in controlled conditions. We hypothesize that OS updates impact computational pipelines mainly through numerical perturbations originating in mathematical libraries, which we simulate using Monte-Carlo arithmetic in a framework called "fuzzy libmath" (FL). We applied this methodology to pre-processing pipelines of the Human Connectome Project, a flagship open-data project in neuroimaging. We found that FL-perturbed pipelines accurately reproduce the variability induced by OS updates and that this similarity is only mildly dependent on simulation parameters. Importantly, we also found between-subject differences were preserved in both cases, though the between-run variability was of comparable magnitude for both FL and OS perturbations. We found the numerical precision in the HCP pre-processed images to be relatively low, with less than 8 significant bits among the 24 available, which motivates further investigation of the numerical stability of components in the tested pipeline. Overall, our results establish that FL accurately simulates results variability due to OS updates, and is a practical framework to quantify numerical uncertainty in neuroimaging.

q-bio.NC↗

arXiv2021

Reducing numerical precision preserves classification accuracy in Mondrian Forests

Mondrian Forests are a powerful data stream classification method, but their large memory footprint makes them ill-suited for low-resource platforms such as connected objects. We explored using reduced-precision floating-point representations to lower memory consumption and evaluated its effect on classification performance. We applied the Mondrian Forest implementation provided by OrpailleCC, a C++ collection of data stream algorithms, to two canonical datasets in human activity recognition: Recofit and Banos \emph{et al}. Results show that the precision of floating-point values used by tree nodes can be reduced from 64 bits to 8 bits with no significant difference in F1 score. In some cases, reduced precision was shown to improve classification performance, presumably due to its regularization effect. We conclude that numerical precision is a relevant hyperparameter in the Mondrian Forest, and that commonly-used double precision values may not be necessary for optimal performance. Future work will evaluate the generalizability of these findings to other data stream classifiers.

cs.LG↗

arXiv2021

An Analysis of Security Vulnerabilities in Container Images for Scientific Data Analysis

Software containers greatly facilitate the deployment and reproducibility of scientific data analyses in various platforms. However, container images often contain outdated or unnecessary software packages, which increases the number of security vulnerabilities in the images, widens the attack surface in the container host, and creates substantial security risks for computing infrastructures at large. This paper presents a vulnerability analysis of container images for scientific data analysis. We compare results obtained with four vulnerability scanners, focusing on the use case of neuroscience data analysis, and quantifying the effect of image update and minification on the number of vulnerabilities. We find that container images used for neuroscience data analysis contain hundreds of vulnerabilities, that software updates remove about two thirds of these vulnerabilities, and that removing unused packages is also effective. We conclude with recommendations on how to build container images with a reduced amount of vulnerabilities.

cs.CR↗

arXiv2021

Modeling the Linux page cache for accurate simulation of data-intensive applications

The emergence of Big Data in recent years has resulted in a growing need for efficient data processing solutions. While infrastructures with sufficient compute power are available, the I/O bottleneck remains. The Linux page cache is an efficient approach to reduce I/O overheads, but few experimental studies of its interactions with Big Data applications exist, partly due to limitations of real-world experiments. Simulation is a popular approach to address these issues, however, existing simulation frameworks do not simulate page caching fully, or even at all. As a result, simulation-based performance studies of data-intensive applications lead to inaccurate results. In this paper, we propose an I/O simulation model that includes the key features of the Linux page cache. We have implemented this model as part of the WRENCH workflow simulation framework, which itself builds on the popular SimGrid distributed systems simulation framework. Our model and its implementation enable the simulation of both single-threaded and multithreaded applications, and of both writeback and writethrough caches for local or network-based filesystems. We evaluate the accuracy of our model in different conditions, including sequential and concurrent applications, as well as local and remote I/Os. We find that our page cache model reduces the simulation error by up to an order of magnitude when compared to state-of-the-art, cacheless simulations.

cs.DC↗

arXiv2020

File-based localization of numerical perturbations in data analysis pipelines

Data analysis pipelines are known to be impacted by computational conditions, presumably due to the creation and propagation of numerical errors. While this process could play a major role in the current reproducibility crisis, the precise causes of such instabilities and the path along which they propagate in pipelines are unclear. We present Spot, a tool to identify which processes in a pipeline create numerical differences when executed in different computational conditions. Spot leverages system-call interception through ReproZip to reconstruct and compare provenance graphs without pipeline instrumentation. By applying Spot to the structural pre-processing pipelines of the Human Connectome Project, we found that linear and non-linear registration are the cause of most numerical instabilities in these pipelines, which confirms previous findings.

q-bio.QM↗

arXiv2020

A benchmark of data stream classification for human activity recognition on connected objects

This paper evaluates data stream classifiers from the perspective of connected devices, focusing on the use case of HAR. We measure both classification performance and resource consumption (runtime, memory, and power) of five usual stream classification algorithms, implemented in a consistent library, and applied to two real human activity datasets and to three synthetic datasets. Regarding classification performance, results show an overall superiority of the HT, the MF, and the NB classifiers over the FNN and the Micro Cluster Nearest Neighbor (MCNN) classifiers on 4 datasets out of 6, including the real ones. In addition, the HT, and to some extent MCNN, are the only classifiers that can recover from a concept drift. Overall, the three leading classifiers still perform substantially lower than an offline classifier on the real datasets. Regarding resource consumption, the HT and the MF are the most memory intensive and have the longest runtime, however, no difference in power consumption is found between classifiers. We conclude that stream learning for HAR on connected objects is challenged by two factors which could lead to interesting future work: a high memory consumption and low F1 scores overall.

cs.LG↗

arXiv2020

Can we Estimate Truck Accident Risk from Telemetric Data using Machine Learning?

Road accidents have a high societal cost that could be reduced through improved risk predictions using machine learning. This study investigates whether telemetric data collected on long-distance trucks can be used to predict the risk of accidents associated with a driver. We use a dataset provided by a truck transportation company containing the driving data of 1,141 drivers for 18 months. We evaluate two different machine learning approaches to perform this task. In the first approach, features are extracted from the time series data using the FRESH algorithm and then used to estimate the risk using Random Forests. In the second approach, we use a convolutional neural network to directly estimate the risk from the time-series data. We find that neither approach is able to successfully estimate the risk of accidents on this dataset, in spite of many methodological attempts. We discuss the difficulties of using telemetric data for the estimation of the risk of accidents that could explain this negative result.

cs.LG↗

arXiv2020

Comparing Perturbation Models for Evaluating Stability of Neuroimaging Pipelines

A lack of software reproducibility has become increasingly apparent in the last several years, calling into question the validity of scientific findings affected by published tools. Reproducibility issues may have numerous sources of error, including the underlying numerical stability of algorithms and implementations employed. Various forms of instability have been observed in neuroimaging, including across operating system versions, minor noise injections, and implementation of theoretically equivalent algorithms. In this paper we explore the effect of various perturbation methods on a typical neuroimaging pipeline through the use of i) targeted noise injections, ii) Monte Carlo Arithmetic, and iii) varying operating systems to identify the quality and severity of their impact. The work presented here demonstrates that even low order computational models such as the connectome estimation pipeline that we used are susceptible to noise. This suggests that stability is a relevant axis upon which tools should be compared, developed, or improved, alongside more commonly considered axes such as accuracy/biological feasibility or performance. The heterogeneity observed across participants clearly illustrates that stability is a property of not just the data or tools independently, but their interaction. Characterization of stability should therefore be evaluated for specific analyses and performed on a representative set of subjects for consideration in subsequent statistical testing. Additionally, identifying how this relationship scales to higher-order models is an exciting next step which will be explored. Finally, the joint application of perturbation methods with post-processing approaches such as bagging or signal normalization may lead to the development of more numerically stable analyses while maintaining sensitivity to meaningful variation.

q-bio.NC↗

arXiv2019

Performance benefits of Intel(R) OptaneTM DC persistent memory for the parallel processing of large neuroimaging data

Open-access neuroimaging datasets have reached petabyte scale, and continue to grow. The ability to leverage the entirety of these datasets is limited to a restricted number of labs with both the capacity and infrastructure to process the data. Whereas Big Data engines have significantly reduced application performance penalties with respect to data movement, their applied strategies (e.g. data locality, in-memory computing and lazy evaluation) are not necessarily practical within neuroimaging workflows where intermediary results may need to be materialized to shared storage for post-processing analysis. In this paper we evaluate the performance advantage brought by Intel(R) OptaneTM DC persistent memory for the processing of large neuroimaging datasets using the two available configurations modes: Memory mode and App Direct mode. We employ a synthetic algorithm on the 76 GiB and 603 GiB BigBrain, as well as apply a standard neuroimaging application on the Consortium for Reliability and Reproducibility (CoRR) dataset using 25 and 96 parallel processes in both cases. Our results show that the performance of applications leveraging persistent memory is superior to that of other storage devices,with the exception of DRAM. This is the case in both Memory and App Direct mode and irrespective of the amount of data and parallelism. Furthermore, persistent memory in App Direct mode is believed to benefit from the use of DRAM as a cache for writing when output data is significantly smaller than available memory. We believe the use of persistent memory will be beneficial to both neuroimaging applications running on HPC or visualization of large, high-resolution images.

cs.PF↗

arXiv2019

High-Resolution Road Vehicle Collision Prediction for the City of Montreal

Road accidents are an important issue of our modern societies, responsible for millions of deaths and injuries every year in the world. In Quebec only, in 2018, road accidents are responsible for 359 deaths and 33 thousands of injuries. In this paper, we show how one can leverage open datasets of a city like Montreal, Canada, to create high-resolution accident prediction models, using big data analytics. Compared to other studies in road accident prediction, we have a much higher prediction resolution, i.e., our models predict the occurrence of an accident within an hour, on road segments defined by intersections. Such models could be used in the context of road accident prevention, but also to identify key factors that can lead to a road accident, and consequently, help elaborate new policies. We tested various machine learning methods to deal with the severe class imbalance inherent to accident prediction problems. In particular, we implemented the Balanced Random Forest algorithm, a variant of the Random Forest machine learning algorithm in Apache Spark. Interestingly, we found that in our case, Balanced Random Forest does not perform significantly better than Random Forest. Experimental results show that 85% of road vehicle collisions are detected by our model with a false positive rate of 13%. The examples identified as positive are likely to correspond to high-risk situations. In addition, we identify the most important predictors of vehicle collisions for the area of Montreal: the count of accidents on the same road segment during previous years, the temperature, the day of the year, the hour and the visibility.

cs.LG↗

arXiv2019

A performance comparison of Dask and Apache Spark for data-intensive neuroimaging pipelines

In the past few years, neuroimaging has entered the Big Data era due to the joint increase in image resolution, data sharing, and study sizes. However, no particular Big Data engines have emerged in this field, and several alternatives remain available. We compare two popular Big Data engines with Python APIs, Apache Spark and Dask, for their runtime performance in processing neuroimaging pipelines. Our evaluation uses two synthetic pipelines processing the 81GB BigBrain image, and a real pipeline processing anatomical data from more than 1,000 subjects. We benchmark these pipelines using various combinations of task durations, data sizes, and numbers of workers, deployed on an 8-node (8 cores ea.) compute cluster in Compute Canada's Arbutus cloud. We evaluate PySpark's RDD API against Dask's Bag, Delayed and Futures. Results show that despite slight differences between Spark and Dask, both engines perform comparably. However, Dask pipelines risk being limited by Python's GIL depending on task type and cluster configuration. In all cases, the major limiting factor was data transfer. While either engine is suitable for neuroimaging pipelines, more effort needs to be placed in reducing data transfer time.

cs.DC↗

arXiv2019

A Conceptual Marketplace Model for IoT Generated Personal Data

We propose a decentralized conceptual marketplace model for IoT generated personal data. Our model is based on a thorough analysis of personal data in a marketplace context, with specific focus on the challenges presented by commercializing IoT generated personal data. Our model introduces a novel perspective on the commercialization of personal data for a marketplace context via risk evaluation and a data licensing framework. We have designed our model to be centered around protecting the privacy and data rights of data generators through model components that effectively assess and modify transaction risks, and formalize transaction agreements by establishing rights of data use and access between buyer and seller. Our model could serve as a blueprint to inform the implementation of a personal data marketplace that respects privacy and ownership.

cs.CY↗

arXiv2019

Evaluation of pilot jobs for Apache Spark applications on HPC clusters

Big Data has become prominent throughout many scientific fields and, as a result, scientific communities have sought out Big Data frameworks to accelerate the processing of their increasingly data-intensive pipelines. However, while scientific communities typically rely on High-Performance Computing (HPC) clusters for the parallelization of their pipelines, many popular Big Data frameworks such as Hadoop and Apache Spark were primarily designed to be executed on dedicated commodity infrastructures. This paper evaluates the benefits of pilot jobs over traditional batch submission for Apache Spark on HPC clusters. Surprisingly, our results show that the speed-up provided by pilot jobs over batch scheduling is moderate to inexistent (0.98 on average) despite the presence of long queuing times. In addition, pilot jobs provide an extra layer of scheduling that complexifies debugging and deployment. We conclude that traditional batch scheduling should remain the default strategy to deploy Apache Spark applications on HPC clusters.

cs.DC↗

arXiv2019

Subject Cross Validation in Human Activity Recognition

K-fold Cross Validation is commonly used to evaluate classifiers and tune their hyperparameters. However, it assumes that data points are Independent and Identically Distributed (i.i.d.) so that samples used in the training and test sets can be selected randomly and uniformly. In Human Activity Recognition datasets, we note that the samples produced by the same subjects are likely to be correlated due to diverse factors. Hence, k-fold cross validation may overestimate the performance of activity recognizers, in particular when overlapping sliding windows are used. In this paper, we investigate the effect of Subject Cross Validation on the performance of Human Activity Recognition, both with non-overlapping and with overlapping sliding windows. Results show that k-fold cross validation artificially increases the performance of recognizers by about 10%, and even by 16% when overlapping windows are used. In addition, we do not observe any performance gain from the use of overlapping windows. We conclude that Human Activity Recognition systems should be evaluated by Subject Cross Validation, and that overlapping windows are not worth their extra computational cost.

cs.LG↗