Issues deriving from the opaque behaviour of prediction-effective, yet non-interpretable, machine learning predictors are being studied and analysed since many decades. One of the main research branches consists of adopting anyway the unintelligible models, thanks to their predictive performance, but queueing to the learning workflow a dedicated technique aimed at post-hoc extracting human-interpretable symbolic knowledge. Following this research line, a growing number of very different knowledge-extraction procedures have been designed over the last four decades, making it difficult for end-users and researches to orient themselves towards the selection of the most suitable one. Accordingly, this survey aims at providing a guide to perform an aware selection of the knowledge-extraction techniques that most probably fit a given task.
Environmental variation shapes acoustic interactions among birds, creating spatial structures in the sonic signature of local species assemblages. Exploring these patterns at regional scales can reveal processes that segregate acoustic strategies along environmental gradients. Here, we examined how the acoustic trait composition of bird assemblages varies at a regional extent in relation to landscape resolution environmental variation. We used data on 2427 bird assemblages and 15 acoustic traits, quantifying the frequency, complexity, rhythm, and duration of vocalisations for 117 species. We used multivariate ordinations to investigate the distribution of species' acoustic traits along climatic and landscape gradients while accounting for spatial and phylogenetic dependencies. We then assessed whether these relationships resulted in directional shifts in the acoustic trait composition of bird assemblages for three key acoustic traits. Our results show that acoustic traits were phylogenetically and spatially clustered and correlated with regional climatic conditions (e.g. lower complexity and isochronous rhythms under higher precipitation and temperature seasonality). Conversely, we found mixed support for the hypothesis that the acoustic signature of species assemblages is shaped by habitat composition within landscapes. For instance, we found urbanisation to be associated with vocalisations featuring broader spectral bandwidths, likely facilitating their propagation under noise pollution, but also greater complexity, which may hinder transmission in urban landscapes. These regional patterns may reflect differences in the structure of acoustic networks within and among species assemblages. Our results thus form a first step towards a regional-level assessment of the environmental and anthropogenic factors that structure or disrupt acoustic connectivity in landscapes.
PurposeThis paper aims to investigate how entrepreneurial small and medium-sized enterprises (SMEs) develop resilience strategies when faced with crises by leveraging interactions within business networks. Using the Industrial Marketing and Purchasing (IMP) perspective, this study explores entrepreneurial SME responses to disruptive events moving towards resilience, robustness and antifragility.Design/methodology/approachThe study adopts a qualitative, abductive approach using six illustrative case studies from diverse sectors across four countries. Data collection involved semi-structured interviews, direct observations, secondary sources and event-based narrative analysis. The study analyses resilience strategies across three interrelated levels, organisational, dyadic and network.FindingsEntrepreneurial SMEs build resilience through six relationally embedded strategies across three levels: (1) Organisational strategies (Frame and Reclaim, Adapt and Advance) manage internal vulnerabilities through data-driven relationship management and business model innovation. (2) Dyadic strategies (Diversify and Thrive, Reconnect and Protect) stabilise relationships through portfolio diversification and emotional capital cultivation. (3) Network strategies (Bridge and Bond, Ally and Amplify) enable collective responses through cross-sector partnerships and institutional alliances. The study finds that the level at which a crisis originates does not determine the level at which firms develop their resilience responses.Practical implicationsThe framework provides entrepreneurial SMEs with strategic options tailored to their network position, resource constraints and crisis type, demonstrating how firms can progress from basic recovery to antifragile growth.Originality/valueTo the best of the authors' knowledge, this study offers the first systematic, multi-level typology of resilience strategies in entrepreneurial SME networks, extending IMP theorising to capture how resilience strategies are co-constructed across organisational, dyadic and network levels.
Nematodes are among the most diverse and abundant animal groups on our planet, and play key roles in ecosystems, agriculture and human health. This review explores their contributions from a One Health perspective, highlighting their impact on environmental, plant, animal and human well‐being. Free‐living nematodes are essential for nutrient cycling, decomposition and microbial regulation, and their responsiveness to various types of environmental perturbations makes them invaluable bioindicators. In agriculture, nematodes play a dual role: plant‐parasitic species pose serious threats to crops, whereas free‐living and entomopathogenic nematodes (EPNs) support sustainable farming by promoting soil health and serving as biocontrol agents. Advances in molecular tools, such as eDNA metabarcoding, have enhanced their use in environmental biomonitoring programmes. In biomedical research, Caenorhabditis elegans has advanced studies on disease, aging and drug discovery, whereas EPNs and marine nematodes show the potential to address antibiotic resistance. This review emphasizes the need for interdisciplinary efforts to fully leverage the ecological, agricultural and biomedical potential of nematodes and to demonstrate their value within the One Health framework in tackling global challenges.
Future experiments at hadron colliders require an evolution of the tracking sensors to ensure sufficient radiation hardness as well as space and time resolution to handle unprecedented particle fluxes. 3D diamond sensors with laser-graphitized electrodes are promising candidates due to their strong binding energy, small atomic number, and high carrier mobility. However, the high resistance of the engraved electrodes delays the propagation of the induced signals towards the readout electronics, thereby degrading the precision of the timing measurements. So far, this effect has been the dominant factor limiting the time resolution of these devices, with other contributions, such as those due to electric field inhomogeneities or electronic noise, typically neglected. Recent advancements in graphitization technology, however, motivate a renewed effort in modeling signal generation in 3D diamond detectors, to achieve more reliable predictions. To this purpose, we apply an extended version of the Ramo-Shockley theorem, describing the effect of signal propagation as a time-dependent weighting potential, obtained by numerically solving the Maxwell's equations in a quasi-static approximation. We developed a custom spectral method solver and validated it against COMSOL MultiPhysics® . The response of the modeled sensor to a beam of particles is then simulated using Garfield++ and is compared to the data acquired in a beam test carried on in 2021 by the TimeSPOT Collaboration at the SPS, at CERN. Based on the results obtained with this simulation workflow, we conclude that reducing the resistivity of the graphitic columns remains the priority for significantly improving the time resolution of 3D diamond detectors. Once achieved, optimization of the detector geometry and readout electronics design will become equally important steps to further enhance the timing performance of these devices.
