Introduction: Transparency of computation is a requirement for assessing the validity of computed results and research claims based upon them; and it is essential for access to, assessment, and reuse of computational components. These components may be subject to methodological or other challenges over time. While reference to archived software and/or data is increasingly common in publications, a single machine-interpretable, integrative representation of how results were derived, that supports defeasible reasoning, has been absent.
Data provenance analysis has been used as an assistive measure for ensuring system integrity. However, such techniques are typically reactive approaches to identify the root cause of an attack in its aftermath. This is in part due to the fact that the collection of provenance metadata often results in a deluge of information that cannot easily be queried and analyzed in real time. This paper presents an approach for proactively reasoning about provenance metadata within the Automatic Cryptographic Data Centric (ACDC) security architecture, a new security infrastructure in which all data interactions are considered at a coarse granularity, similar to the Function as a Service model. At this scale, we have found that data interactions are manageable for the proactive specification and evaluation of provenance policies—constraints placed on provenance metadata to prevent the consumption of untrusted data. This paper provides a model for proactively evaluating provenance metadata in the ACDC paradigm as well as a case study of an electronic voting scheme to demonstrate the applicability of ACDC and the provenance policies needed to ensure data integrity.
Software repositories contain information about source code, software development processes, and team interactions. We combine the provenance of development processes with code security analysis results to provide fast feedback on the software’s design and security issues. Results from queries of the provenance graph drives the security analysis, which are conducted on certain events—such as commits or pull requests by external contributors. We evaluate our method on Open Source projects that are developed under time pressure and use Germany’s COVID-19 contact tracing app ‘Corona-Warn-App’ as a case study.
We present here a provenance management system adapted to astronomical projects needs. We collected use cases from various astronomy projects and defined a data model in the ecosystem developed by the IVOA (International Virtual Observatory Alliance). From those use cases, we observed that some projects already have data collections generated and archived, from which the provenance has to be extracted (provenance "on top"), and some projects are building complex pipelines that automatically capture provenance information during the data processing (capture "inside"). Different tools and prototypes have been developed and tested to capture, store, access and visualize the provenance information, which participate to the shaping of a full provenance management system able to handle detailed provenance information.
The duration of the life cycle in deep neural networks (DNN) depends on the data configuration decisions that lead to success in obtaining models. Analyzing hyperparameters along the evolution of the network’s execution allows for adapting the data. Provenance data derivation traces help the parameter fine-tuning by providing a global data picture with clear dependencies. Provenance can also contribute to the interpretation of models resulting from the DNN life cycle. However, there are challenges in collecting hyperparameters and in modeling the relationships between the data involved in the DNN life cycle to build a provenance database. Current approaches adopt different notions of provenance in their representation and require the execution of the DNN under a specific software framework, which limits interoperability and flexibility when choosing the DNN execution environment. This work presents a provenance data-based approach to address these challenges, proposing a collection mechanism with flexibility in the choice and representation of data to be analyzed. Experiments of the approach, using a convolutional neural network focused on image recognition, provide evidence of the flexibility, the efficiency of data collection, the analysis and the validation of network data.
This proposed demonstration introduces ProvViz [9], an intuitive in-browser PROV editor and visualiser that targets both users familiar and unfamiliar with the syntax of PROV documents. The layout of the editor is composed of a text-editor and an interactive visualiser which is also released as a standalone UI component on the Node Package Manager (NPM) registry ( https://npmjs.org/provviz ).
In the originally published version, the DOI in Reference 7 “Wittner, R., et al.: EOSClife common provenance model. EOSC-Life deliverable D6.2 (2021)” on p.225 was missing. The DOI “10.5281/zenodo.4705074” has been added.
Industrial Internet of Things (IoT) systems are becoming increasingly complex with interconnected field devices. To cope with the complexity and to facilitate further interoperability between heterogeneous field devices, it is necessary to have a holistic understanding of different types of provenance information about field devices and associated processes. In this paper, we investigate the potential integration of such provenance with the W3C Web of Things and current industrial standards to realize more explainable, efficient, and safer industrial IoT systems.
Whilst the need to record and understand the evolution of data, together with the processes and users associated with those changes, is now widely appreciated, the uptake of solutions to these issues remains slow. Data provenance techniques have the potential to provide such an understanding, but their use is often considered a specialist activity, requiring detailed knowledge of standards such as W3C PROV. In this work, we introduce ProvIt, a suite of tools designed to lower the barriers to entry for the use of provenance technology. We demonstrate the utility of ProvIt by using it to add provenance capabilities to the Jupyter IDE, in order to provide insight into the tools used by a group of researchers analysing a COVID-19 dataset.
Computational notebooks allow users to persist code, results, and explanations together, making them important artifacts in understanding research. However, these notebooks often do not record the full provenance of results because steps can be repeated, reordered, or removed. This can lead to inconsistencies between what the authors found and recorded, and what others see when they attempt to examine those results. However, these notebooks do offer some clues that help us infer and understand what may have happened. This paper presents techniques to unearth patterns and develop hypotheses about how the original results were obtained. The work uses statistics from a large corpora of notebooks to build the probable provenance of a notebook’s state. Results show these techniques can help others understand notebooks that may have been archived without proper preservation.
Provenance is information about entities, activities, and people involved in producing a piece of data or a thing, which can be used to form assessments about the data or the thing’s quality, reliability, or trustworthiness. PROV-DM is the conceptual data model that forms the basis for the W3C provenance (PROV) family of specifications. In this paper, we propose a new serialization for PROV in JSON called PROV-JSONLD. It provides a lightweight representation of PROV expressions in JSON, which is suitable to be processed by Web applications, while maintaining a natural encoding that is familiar with PROV practitioners. In addition, PROV-JSONLD exploits JSON-LD to define a semantic mapping that conforms to the PROV-O specification and, hence, the encoded PROV expressions can be readily processed as Linked Data. Finally, we show that the serialization is also efficiently processable in our evaluation. Overall, PROV-JSONLD is designed to be suitable for interchanging provenance information in Web and Linked Data applications, to offer a natural encoding of provenance for its targeted audience, and to allow for fast processing.
Curated scientific databases play an important role in the scientific endeavour and support is needed for the significant effort that goes into their creation and maintenance. This demonstration and case study illustrate how curation support has been developed in the Links cross-tier programming language, a functional, strongly typed language with language-integrated query and support for temporal databases. The chosen case study uses weekly released Covid-19 fatality figures from the Scottish government which exhibit updates to previously released data. This data allows the capture and query of update provenance in our prototype. This demonstration will highlight the potential for language-integrated support for curation to simplify and streamline prototyping of web-applications in support of scientific databases
Provenance templates are an established methodology for the capture of provenance data. Each template defines the provenance of a domain-specific action in abstract form, which may then be instantiated as required by a single call to a given service interface. This approach, whilst simplifying the process of recording provenance for the user, introduces computational and storage demands on the capture process, particularly when used by clients with write-intensive provenance requirements such as other service-based software. To address these issues, we adopt a new approach based upon delayed instantiation and present a revised, two-part paradigm for template-driven provenance, in which we separate capture and query functionality to improve the overall efficiency of the model. A dedicated capture service is first employed to record template service requests in a relational database in the form of a meta-level description detailing the construction of each document. These low-overhead records are then accessed by an independent query service to construct views of concrete provenance documents for specific time frames as and when required by the user. These views may subsequently be analysed using query templates, a new technique defined here whereby templates can also be used to search for any matching subgraphs within a document and return the respective instantiating substitutions. We evaluate the performance gains of our new system in the context of Phenoflow, an electronic health record (EHR) phenotyping platform.
Data cleaning is an essential component of data preparation in machine learning and other data science workflows, and is widely recognized as the most time-consuming and error-prone part when working with real-world data. How data was prepared and cleaned has a significant impact on the reliability and trustworthiness of results of any subsequent analysis. Transparent data cleaning not only requires that provenance (i.e., operation history and value changes) be captured, but also that those changes are easy to explore and evaluate: The data scientists who prepare the data, as well as others who want to reuse the cleaned data for their studies, need to be able to easily explore and query its data cleaning history. We have developed a domain-specific provenance model for data cleaning that supports the kind of provenance questions that data scientists need to answer when inspecting and debugging data preparation histories. The design of the model was driven by the need (i) to answer relevant, user-oriented provenance questions, and (ii) to do so in an effective and efficient manner. The model is a refinement of an earlier provenance model and has been implemented as a companion tool to OpenRefine, a popular, open source tool for data cleaning.
Long-running applications nowadays are increasingly instrumented to continuously log provenance. In that context, we observe an emerging need for processing fragments of provenance continuously produced by applications. Thus, there is an increasing requirement for processing of provenance incrementally, while the application is still running, to replace batch processing of a complete provenance dataset available only after the application has completed. A type of processing of particular interest is summarising provenance graphs, which has been proposed as an effective way of extracting key features of provenance and storing them in an efficient manner. To that goal, summarisation makes use of provenance types, which, in loose terms, are an encoding of the neighbourhood of nodes. This paper shows that the process of creating provenance summaries of continuously provided data can benefit from a mode of incremental processing of provenance types. We also introduce the concept of a library of types to reduce the need for storing copies of the same string representations for types multiple times. Further, we show that the computational complexity associated with the task of inferring types is, in most common cases, the best possible: only new nodes have to be processed. We also identify and analyse the exception scenarios. Finally, although our library of types, in theory, can be exponentially large, we present empirical results that show it is quite compact in practice.
We propose a novel method of assessing OpenStreetMap data using the concept of Data Maturity. Based on research into data quality and trust in user generated content, this is a set of measurements that can be derived from provenance data extracted from OpenStreetMap edit history.
Users typically interact with a database by asking queries and examining the results. We refer to the user examining the query results and asking follow-up questions as query result exploration. Our work builds on two decades of provenance research useful for query result exploration. Three approaches for computing provenance have been described in the literature: lazy, eager, and hybrid. We investigate lazy and eager approaches that utilize constraints that we have identified in the context of query result exploration, as well as novel hybrid approaches. For the TPC-H benchmark, these constraints are applicable to 19 out of the 22 queries, and result in a better performance for all queries that have a join. Furthermore, the performance benefits from our approaches are significant, sometimes several orders of magnitude.
Data preparation modules are ubiquitous and are used to perform, amongst other things, operations such as record retrieval, format transformation, data combination to name a few. To assist scientists in the task of discovering suitable modules, semantic annotations can be leveraged. Experience suggests, however, that while such annotations are useful in describing the inputs and outputs of a module, they fail in crisply describing the functionality performed by the module. To overcome this issue, we outline in this poster paper a solution that utilizes semantic annotations describing the inputs and outputs of modules together with data examples that characterize modules’ behavior as ingredients for querying data preparation modules. Data examples are constructed using retrospective provenance of module executions. The discovery strategy that we devised is iterative in that it allows scientists to explore existing modules by providing feedback on data examples.
