Abstract For a long time now, datasets containing scientific articles have been crucial to the analysis and recognition of document images. These document collections have frequently served as a testing ground for cutting-edge methods for optical character recognition, layout analysis, and document understanding in general. We thoroughly analyze and compare many datasets proposed for layout analysis of scientific documents, ranging from small collections of scanned papers to modern large-scale datasets containing digital-born papers, which have been proposed to train deep learning-based methods. Furthermore, we outline a detailed taxonomy of the annotation procedures used considering manual, automatic, and generative approaches, and we analyze their benefits and drawbacks. This survey is meant to provide the reader with a review of the most used benchmarks together with detailed information on data, annotations, and complexity, helping scholars to identify the most suitable dataset for their tasks of interest. We also discuss possible open problems to further enhance datasets to support research in the layout analysis of scientific articles.

Quantitative evaluation of stochastic models supports early verification of design choices and assessment of non-functional requirements. Model Driven Engineering (MDE) leverages automated derivation of formal stochastic models from semi-formal artifacts of the Unified Modeling Language (UML) to facilitate deployment of quantitative evaluation methods without disrupting industrial practices. As a major limitation, when generally distributed (GEN) temporal parameters are considered to enhance the model expressivity, the structure and complexity of the underlying stochastic process cannot be easily controlled, possibly impairing the model analyzability. We present a hierarchical modeling formalism based on UML statecharts with GEN durations, designed to guarantee ease of modeling and efficient evaluation of steady-state or transient behaviour until absorption. To this end, fairly lax restrictions are applied to the model syntax to enable separate analysis of the Semi-Markov Process (SMP) underlying each model component. Scalability of solution is assessed by analyzing a suite of synthetic models referred to the context of timed Failure Logic Analysis (FLA) of component-based systems, specifically designed to point out each factor of computational complexity. Notably, the analysis derives both the probability that the system is in each step before failure and the Cumulative Distribution Function (CDF) of the duration of the overall failure process. A challenging case study that significantly and jointly stresses the main factors of computational complexity is finally addressed, performing steady-state analysis of a non-Markovian variant of a server virtualized system from the literature on software rejuvenation.

Runtime predictive analysis of quantitative models can support software reliability in various application scenarios. The spread of logging technologies promotes approaches where such models are learned from observed events. We consider a system visiting transient states of a hidden process until reaching a final state and producing observations with stochastic arrival times and types conditioned by visited states, and we abstract it as a marked Markov modulated Poisson Process (MMMPP) with left-to right structure. We present an Expectation-Maximization (EM) algorithm that learns the MMMPP parameters from observation sequences acquired in repeated execution of the transient behavior, and we use the model at runtime to infer the current state of the process from actual observed events and to dynamically evaluate the remaining time to the final state. The approach is illustrated using synthetic datasets generated from a stochastic attack tree of the literature enriched with an observation model associating each state with an expected statistics of observation types and arrival times. Accuracy of prediction is evaluated under different variability of hidden states sojourn durations and of the observations arrival process, and compared against previous literature that mainly exploits either the timing or the types of observed events.