Stream macroinvertebrate communities in restored and impacted catchments respond differently to climate, land-use, and runoff over a decade

Identifying which environmental drivers underlie degradation and improvements of ecological communities is a fundamental goal of ecology. Achieving this goal is a challenge due to diverse trends in both environmental conditions and ecological communities across regions, and it is constrained by the lack of long-term parallel monitoring of environmental and community data needed to study causal relationships. Here, we identify key environmental drivers using a high-resolution environmental - ecological dataset, an ensemble of the Soil and Water Assessment Tool (SWAT+) model, and ecological models to investigate effects of climate, land-use, and runoff on the decadal trend (2012-2021) of stream macroinvertebrate communities in a restored urban catchment and an impacted catchment with mixed land-uses in Germany. The decadal trends showed decreased precipitation, increased temperature, and reduced anthropogenic land-uses, which led to opposing runoff trends - with decreased runoff in the restored catchment and increased runoff in the impacted catchment. The two catchments also varied in decadal trends of taxonomic and trait composition and metrics. The most significant improvements over time were recorded in communities of the restored catchment sites, which have become wastewater free since 2007 to 2009. Within the restored catchment sites, community metric trends were primarily explained by land-use and evaporation trends, while community composition trends were mostly associated with precipitation and runoff trends. Meanwhile, the communities in the impacted catchment did not undergo significant changes between 2012 and 2021, likely influenced by the effects of prolonged droughts following floods after 2018. The results of our study confirm the significance of restoration and land-use management in fostering long-term improvements in stream communities, while climate change remains a prodigious threat. The coupling of long-term biodiversity monitoring with concurrent sampling of relevant environmental drivers is critical for preventative and restorative management in ecology.

How tolerances, competition and dispersal shape benthic invertebrate colonisation in restored urban streams

Biotic communities often respond poorly to river restoration activities and the drivers of community recovery after restoration are not fully understood. According to the Asymmetric Response Concept (ARC), dispersal capacity, species tolerances to stressors, and biotic interactions are three key drivers influencing community recovery of restored streams. However, the ARC remains to be tested. Here we used a dataset on benthic invertebrate communities of eleven restored stream sections in a former open sewer system that were sampled yearly over a period of eleven years. We applied four indices that reflect tolerance against chloride and organic pollution, the community's dispersal capacity and strength of competition to the benthic invertebrate taxa lists of each year and site. Subsequently, we used generalised linear mixed models to analyse the change of these indices over time since restoration. Dispersal capacity was high directly after restoration but continuously decreased over time. The initial communities thus consisted of good dispersers and were later joined by more slowly dispersing taxa. The tolerance to organic pollution also decreased over time, reflecting continuous improvement of water quality and an associated increase of sensitive species. On the contrary, chloride tolerances did not change, which could indicate a stable chloride level throughout the sampling period. Lastly, competition within the communities, reflected by interspecific trait niche overlap, increased with time since restoration. We show that recovery follows a specific pattern that is comparable between sites. Benthic communities change from tolerant, fast dispersing generalists to more sensitive, slowly dispersing specialists exposed to stronger competition. Our results lay support to the ARC (increasing role of competition, decreasing role of dispersal) but also underline that certain tolerances may still shape communities a decade after restoration. Disentangling the drivers of macroinvertebrate colonisation can help managers to better understand recovery trajectories and to define more realistic restoration targets.

Bending Behavior of Hybrid Timber-Steel Beams

Driven by climate change and the need for a more sustainable construction sector, policy is increasingly demanding and promoting timber hybrid construction methods. In the German state of Baden-Württemberg, every new public building has to be of timber or timber hybrid construction (Holzbauoffensive BW). The objective of multi-story buildings with large floor spans can only be achieved in a resource-efficient way by hybrid constructions combining timber and steel components. A research project recently completed at the Karlsruhe Institute of Technology was aimed at the development and systematic investigation of hybrid bending beams in which an advantageous combination of the materials steel and timber is used. For this purpose, steel profiles are integrated into timber cross-sections in a shear-resistant manner by adhesive bonding. As part of the experimental, numerical and analytical investigations, different cross-sections of steel and timber, as well as different construction materials, were considered (GL24h, LVL48p, LVL80p, S355 and S420). The results of large-scale four-point bending tests illustrate the potential of this new hybrid construction method. Depending on the geometry and material combinations tested, the bending stiffness could be increased by up to 250%, and the load-carrying capacity by up to 120%, compared to a glulam beam with identical dimensions.

Multi-decadal improvements in the ecological quality of European rivers are not consistently reflected in biodiversity metrics

Humans impact terrestrial, marine and freshwater ecosystems, yet many broad-scale studies have found no systematic, negative biodiversity changes (for example, decreasing abundance or taxon richness). Here we show that mixed biodiversity responses may arise because community metrics show variable responses to anthropogenic impacts across broad spatial scales. We first quantified temporal trends in anthropogenic impacts for 1,365 riverine invertebrate communities from 23 European countries, based on similarity to least-impacted reference communities. Reference comparisons provide necessary, but often missing, baselines for evaluating whether communities are negatively impacted or have improved (less or more similar, respectively). We then determined whether changing impacts were consistently reflected in metrics of community abundance, taxon richness, evenness and composition. Invertebrate communities improved, that is, became more similar to reference conditions, from 1992 until the 2010s, after which improvements plateaued. Improvements were generally reflected by higher taxon richness, providing evidence that certain community metrics can broadly indicate anthropogenic impacts. However, richness responses were highly variable among sites, and we found no consistent responses in community abundance, evenness or composition. These findings suggest that, without sufficient data and careful metric selection, many common community metrics cannot reliably reflect anthropogenic impacts, helping explain the prevalence of mixed biodiversity trends.

Understanding temporal variability across trophic levels and spatial scales in freshwater ecosystems

A tenet of ecology is that temporal variability in ecological structure and processes tends to decrease with increasing spatial scales (from locales to regions) and levels of biological organization (from populations to communities). However, patterns in temporal variability across trophic levels and the mechanisms that produce them remain poorly understood. Here we analyzed the abundance time series of spatially structured communities (i.e., metacommunities) spanning basal resources to top predators from 355 freshwater sites across three continents. Specifically, we used a hierarchical partitioning method to disentangle the propagation of temporal variability in abundance across spatial scales and trophic levels. We then used structural equation modeling to determine if the strength and direction of relationships between temporal variability, synchrony, biodiversity, and environmental and spatial settings depended on trophic level and spatial scale. We found that temporal variability in abundance decreased from producers to tertiary consumers but did so mainly at the local scale. Species population synchrony within sites increased with trophic level, whereas synchrony among communities decreased. At the local scale, temporal variability in precipitation and species diversity were associated with population variability (linear partial coefficient, β = 0.23) and population synchrony (β = -0.39) similarly across trophic levels, respectively. At the regional scale, community synchrony was not related to climatic or spatial predictors, but the strength of relationships between metacommunity variability and community synchrony decreased systematically from top predators (β = 0.73) to secondary consumers (β = 0.54), to primary consumers (β = 0.30) to producers (β = 0). Our results suggest that mobile predators may often stabilize metacommunities by buffering variability that originates at the base of food webs. This finding illustrates that the trophic structure of metacommunities, which integrates variation in organismal body size and its correlates, should be considered when investigating ecological stability in natural systems. More broadly, our work advances the notion that temporal stability is an emergent property of ecosystems that may be threatened in complex ways by biodiversity loss and habitat fragmentation.

The recovery of European freshwater biodiversity has come to a halt

Owing to a long history of anthropogenic pressures, freshwater ecosystems are among the most vulnerable to biodiversity loss1. Mitigation measures, including wastewater treatment and hydromorphological restoration, have aimed to improve environmental quality and foster the recovery of freshwater biodiversity2. Here, using 1,816 time series of freshwater invertebrate communities collected across 22 European countries between 1968 and 2020, we quantified temporal trends in taxonomic and functional diversity and their responses to environmental pressures and gradients. We observed overall increases in taxon richness (0.73% per year), functional richness (2.4% per year) and abundance (1.17% per year). However, these increases primarily occurred before the 2010s, and have since plateaued. Freshwater communities downstream of dams, urban areas and cropland were less likely to experience recovery. Communities at sites with faster rates of warming had fewer gains in taxon richness, functional richness and abundance. Although biodiversity gains in the 1990s and 2000s probably reflect the effectiveness of water-quality improvements and restoration projects, the decelerating trajectory in the 2010s suggests that the current measures offer diminishing returns. Given new and persistent pressures on freshwater ecosystems, including emerging pollutants, climate change and the spread of invasive species, we call for additional mitigation to revive the recovery of freshwater biodiversity.

Multidecadal data indicate increase of aquatic insects in Central European streams

In recent years, declining insect biodiversity has sparked interest among scientists and drawn the attention of society and politicians. However, our understanding of the extent of this decline is incomplete, particularly for freshwater insects that provide a key trophic link between aquatic and terrestrial ecosystems, but that are also especially vulnerable to climate change. To investigate the response of freshwater insects to climate change, we quantified shifts in insect abundance and diversity across 7264 samples covering Central Europe during 1990-2018 and related these changes to annual data on temperature and precipitation. We observed both increases in richness (10.6 %) and abundance (9.5 %) of freshwater insects over the past three decades. These changes were related to increases in summer temperature and summer precipitation, which had negative effects on species richness, and to increases in winter temperature and precipitation, which had positive effects. Further we found that increased temperature was generally related to increased abundance, whereas increased precipitation was associated with declines, thus highlighting the particularly varying impacts on differing insect orders. Given that freshwater insects have been more severely affected by global change than marine and terrestrial species, the observed increases are a positive sign, but the overall situation of freshwater invertebrates is still critical.

Long-term trends and drivers of biological invasion in Central European streams

Rates of biological invasion continue to accelerate and threaten the structure and function of ecosystems worldwide. High habitat connectivity, multiple pathways, and inadequate monitoring have rendered aquatic ecosystems vulnerable to species introductions. Past riverine invasion dynamics were largely restricted to large rivers, leaving out smaller rivers that commonly harbour high freshwater biodiversity. Moreover, biodiversity time series have rarely been used to investigate invasions across larger spatial-temporal scales, limiting our understanding of aquatic invasion dynamics. Here, we used 6067 benthic invertebrate samples from streams and small rivers from the EU Water Framework Directive monitoring program collected across Central Europe between 2000 and 2018 to assess temporal changes to benthic invertebrate communities as well as non-native species. We assessed invasion rates according to temperature, precipitation, elevation, latitude, longitude, and stream type. Overall, average daily temperatures significantly increased by 0.02 °C per annum (0.34 °C in total) while annual precipitation significantly decreased by 0.01 mm per annum (-67.8 mm over the study period), paralleled with significant increases in overall species richness (12.3 %) and abundance (14.9 %); water quality was relatively stable. Non-native species richness increased 5-fold and abundance 40-fold, indicating an ongoing community shift from native to non-native species. The observed increase in invasions was stronger in low mountain rivers compared to low mountain streams, with the share of non-native species abundance and richness declining with increasing elevation and latitude but increasing with temperature. We found thermophilic non-native species invasion success was greatest in larger sized streams, at lower latitudes, lower elevations and higher temperatures. These results indicate that widespread environmental characteristics (i.e., temperature) could heighten invasion success and confer refuge effects (i.e., elevation and latitude) in higher sites. High altitude and latitude environments should be prioritised for prevention efforts, while biosecurity and management should be improved in lowland areas subject to greater anthropogenic pressure, where non-native introductions are more likely.

Taxonomic and functional reorganization in Central European stream macroinvertebrate communities over 25 years

Climate warming can lead to a replacement of species that favour cold temperatures by species that favour warm temperatures. However, the implications of such thermic shifts for the functioning of ecosystems remain poorly understood. Here, we used stream macroinvertebrate biological and ecological traits to quantify the relative contribution of cold, intermediate and warm temperature-adapted taxa to changes in community functional diversity (FD) using a dataset of 3781 samples collected in Central Europe over 25 years, from 1990 to 2014. Our analyses indicated that functional diversity of stream macroinvertebrate communities increased over the study period. This gain was driven by a net 39 % increase in the richness of taxa that favour intermediate temperatures, which comprise the highest share in the community, and to a 97 % increase in the richness of taxa that favour warm temperatures. These warm temperature-adapted taxa displayed a distinct and more diverse suite of functional traits compared to the cold temperature-adapted group and thus contributed disproportionately to local FD on a per-taxon basis. At the same time, taxonomic beta-diversity declined significantly within each thermal group, in association with increasing local taxon richness. This study shows that over recent decades, small low-mountain streams in Central Europe have experienced a process of thermophilization and increasing functional diversity at local scales. However, a progressive homogenisation occurred at the regional scale, with communities converging towards similar taxonomic composition. As the reported increase in local functional diversity can be attributed mostly to the intermediate temperature-adapted taxa and a few expanding warm temperature-adapted taxa, these patterns could mask more subtle loss of sensitive cold temperature-adapted taxa with irreplaceable functional traits. In light of increasing climate warming, preservation of cold habitat refuges, should be considered a priority in river conservation.

The Asymmetric Response Concept explains ecological consequences of multiple stressor exposure and release

Our capacity to predict trajectories of ecosystem degradation and recovery is limited, especially when impairments are caused by multiple stressors. Recovery may be fast or slow and either complete or partial, sometimes result in novel ecosystem states or even fail completely. Here, we introduce the Asymmetric Response Concept (ARC) that provides a basis for exploring and predicting the pace and magnitude of ecological responses to, and release from, multiple stressors. The ARC holds that three key mechanisms govern population, community and ecosystem trajectories. Stress tolerance is the main mechanism determining responses to increasing stressor intensity, whereas dispersal and biotic interactions predominantly govern responses to the release from stressors. The shifting importance of these mechanisms creates asymmetries between the ecological trajectories that follow increasing and decreasing stressor intensities. This recognition helps to understand multiple stressor impacts and to predict which measures will restore communities that are resistant to restoration.

Long-term trends in abundances of non-native species across biomes, realms, and taxonomic groups in Europe

Rates of biological invasion have increased over recent centuries and are expected to increase in the future. Whereas increasing rates of non-native species incursions across realms, taxonomic groups, and regions are well-reported, trends in abundances within these contexts have lacked analysis due to a paucity of long-term data at large spatiotemporal scales. These knowledge gaps impede prioritisation of realms, regions, and taxonomic groups for management. We analysed 180 biological time series (median 15 ± 12.8 sampling years) mainly from Long-Term Ecological Research (LTER) sites comprising abundances of marine, freshwater, and terrestrial non-native species in Europe. A high number (150; 83,3 %) of these time series were invaded by at least one non-native species. We tested whether (i) local long-term abundance trends of non-native species are consistent among environmental realms, taxonomic groups, and regions, and (ii) if any detected trend can be explained by climatic conditions. Our results indicate that abundance trends at local scales are highly variable, with evidence of declines in marine and freshwater long-term monitoring sites, despite non-native species reports increasing rapidly since the late 1970s. These declines were driven mostly by abundance trends in non-native fish, birds, and invertebrate species in three biogeographic regions (Continental, Atlantic, and the North Sea). Temperature and precipitation were important predictors of observed abundance trends across Europe. Yet, the response was larger for species with already declining trends and differed among taxa. Our results indicate that trends in biological invasions, especially across different taxonomic groups, are context-dependent and require robust local data to understand long-term trends across contexts at large scales. While the process of biological invasion is spatiotemporally broad, economic or ecological impacts are generally realised on the local level. Accordingly, we urge proactive and coordinated management actions from local to large scales, as invasion impacts are substantial and dynamics are prone to change.

Alien species and climate change drive shifts in a riverine fish community and trait compositions over 35 years

Alien fish substantially impact aquatic communities. However, their effects on trait composition remain poorly understood, especially at large spatiotemporal scales. Here, we used long-term biomonitoring data (1984-2018) from 31 fish communities of the Rhine river in Germany to investigate compositional and functional changes over time. Average total community richness increased by 49 %: it was stable until 2004, then declined until 2010, before increasing until 2018. Average abundance decreased by 9 %. Starting from 198 individuals/m² in 1984 abundance largely declined to 23 individuals/m² in 2010 (-88 %), and then consequently increased by 678 % up to 180 individuals/m² until 2018. Increases in abundance and richness starting around 2010 were mainly driven by the establishment of alien species: while alien species represented 5 % of all species and 0.1 % of total individuals in 1993, it increased to 30 % (7 species) and 32 % of individuals in 2018. Concomitant to the increase in alien species, average native species richness and abundance declined by 26 % and 50 % respectively. We identified increases in temperature, precipitation, abundance and richness of alien fish driving compositional changes after 2010. To get more insights on the impacts of alien species on fish communities, we used 12 biological and 13 ecological traits to compute four trait metrics each. Ecological trait dispersion increased before 2010, probably due to diminishing ecologically similar native species. No changes in trait metrics were measured after 2010, albeit relative shares of expressed trait modalities significantly changing. The observed shift in trait modalities suggested the introduction of new species carrying similar and novel trait modalities. Our results revealed significant changes in taxonomic and trait compositions following alien fish introductions and climatic change. To conclude, our analyses show taxonomic and functional changes in the Rhine river over 35 years, likely indicative of future changes in ecosystem services.

Evaluating the biological validity of European river typology systems with least disturbed benthic macroinvertebrate communities

Humans have severely altered freshwater ecosystems globally, causing a loss of biodiversity. Regulatory frameworks, like the Water Framework Directive, have been developed to support actions that halt and reverse this loss. These frameworks use typology systems that summarize freshwater ecosystems into environmentally delineated types. Within types, ecosystems that are minimally impacted by human activities, i.e., in reference conditions, are expected to be similar concerning physical, chemical, and biological characteristics. This assumption is critical when water quality assessments rely on comparisons to type-specific reference conditions. Lyche Solheim et al. (2019) developed a pan-European river typology system, the Broad River Types, that unifies the national Water Framework Directive typology systems and is gaining traction within the research community. However, it is unknown how similar biological communities are within these individual Broad River Types. We used analysis of similarities and classification strength analysis to examine if the Broad River Types delineate distinct macroinvertebrate communities across Europe and whether they outperform two ecoregional approaches: the European Biogeographical Regions and Illies' Freshwater Ecoregions. We determined indicator and typical taxa for the types of all three typology systems and evaluated their distinctiveness. All three typology systems captured more variation in macroinvertebrate communities than random combinations of sites. The results were similar among typology systems, but the Broad River Types always performed worse than either the Biogeographic Regions or Illies' Freshwater Ecoregions. Despite reaching statistical significance, the statistics of analysis of similarity and classification strength were low in all tests indicating substantial overlap among the macroinvertebrate communities of different types. We conclude that the Broad River Types do not represent an improvement upon existing freshwater typologies when used to delineate macroinvertebrate communities and we propose future avenues for advancement: regionally constrained types, better recognition of intermittent rivers, and consideration of biotic communities.

Invasion impacts and dynamics of a European-wide introduced species

Globalization has led to the introduction of thousands of alien species worldwide. With growing impacts by invasive species, understanding the invasion process remains critical for predicting adverse effects and informing efficient management. Theoretically, invasion dynamics have been assumed to follow an "invasion curve" (S-shaped curve of available area invaded over time), but this dynamic has lacked empirical testing using large-scale data and neglects to consider invader abundances. We propose an "impact curve" describing the impacts generated by invasive species over time based on cumulative abundances. To test this curve's large-scale applicability, we used the data-rich New Zealand mud snail Potamopyrgus antipodarum, one of the most damaging freshwater invaders that has invaded almost all of Europe. Using long-term (1979-2020) abundance and environmental data collected across 306 European sites, we observed that P. antipodarum abundance generally increased through time, with slower population growth at higher latitudes and with lower runoff depth. Fifty-nine percent of these populations followed the impact curve, characterized by first occurrence, exponential growth, then long-term saturation. This behaviour is consistent with boom-bust dynamics, as saturation occurs due to a rapid decline in abundance over time. Across sites, we estimated that impact peaked approximately two decades after first detection, but the rate of progression along the invasion process was influenced by local abiotic conditions. The S-shaped impact curve may be common among many invasive species that undergo complex invasion dynamics. This provides a potentially unifying approach to advance understanding of large-scale invasion dynamics and could inform timely management actions to mitigate impacts on ecosystems and economies.

Sampling and modelling rare species: Conceptual guidelines for the neglected majority

Biodiversity conservation faces a methodological conundrum: Biodiversity measurement often relies on species, most of which are rare at various scales, especially prone to extinction under global change, but also the most challenging to sample and model. Predicting the distribution change of rare species using conventional species distribution models is challenging because rare species are hardly captured by most survey systems. When enough data are available, predictions are usually spatially biased towards locations where the species is most likely to occur, violating the assumptions of many modelling frameworks. Workflows to predict and eventually map rare species distributions imply important trade-offs between data quantity, quality, representativeness and model complexity that need to be considered prior to survey and analysis. Our opinion is that study designs need to carefully integrate the different steps, from species sampling to modelling, in accordance with the different types of rarity and available data in order to improve our capacity for sound assessment and prediction of rare species distribution. In this article, we summarize and comment on how different categories of species rarity lead to different types of occurrence and distribution data depending on choices made during the survey process, namely the spatial distribution of samples (where to sample) and the sampling protocol in each selected location (how to sample). We then clarify which species distribution models are suitable depending on the different types of distribution data (how to model). Among others, for most rarity forms, we highlight the insights from systematic species-targeted sampling coupled with hierarchical models that allow correcting for overdispersion and spatial and sampling sources of bias. Our article provides scientists and practitioners with a much-needed guide through the ever-increasing diversity of methodological developments to improve the prediction of rare species distribution depending on rarity type and available data.

Global maps of soil temperature

Research in global change ecology relies heavily on global climatic grids derived from estimates of air temperature in open areas at around 2 m above the ground. These climatic grids do not reflect conditions below vegetation canopies and near the ground surface, where critical ecosystem functions occur and most terrestrial species reside. Here, we provide global maps of soil temperature and bioclimatic variables at a 1-km2 resolution for 0-5 and 5-15 cm soil depth. These maps were created by calculating the difference (i.e. offset) between in situ soil temperature measurements, based on time series from over 1200 1-km2 pixels (summarized from 8519 unique temperature sensors) across all the world's major terrestrial biomes, and coarse-grained air temperature estimates from ERA5-Land (an atmospheric reanalysis by the European Centre for Medium-Range Weather Forecasts). We show that mean annual soil temperature differs markedly from the corresponding gridded air temperature, by up to 10°C (mean = 3.0 ± 2.1°C), with substantial variation across biomes and seasons. Over the year, soils in cold and/or dry biomes are substantially warmer (+3.6 ± 2.3°C) than gridded air temperature, whereas soils in warm and humid environments are on average slightly cooler (-0.7 ± 2.3°C). The observed substantial and biome-specific offsets emphasize that the projected impacts of climate and climate change on near-surface biodiversity and ecosystem functioning are inaccurately assessed when air rather than soil temperature is used, especially in cold environments. The global soil-related bioclimatic variables provided here are an important step forward for any application in ecology and related disciplines. Nevertheless, we highlight the need to fill remaining geographic gaps by collecting more in situ measurements of microclimate conditions to further enhance the spatiotemporal resolution of global soil temperature products for ecological applications.

Decline in niche specialization and trait β-diversity in benthic invertebrate communities of Central European low-mountain streams over 25 years

Biotic homogenization is one of the key aspects of the current biodiversity crisis. Here we analyzed the trends of three facets of niche homogenization, i.e. niche specialization, trait α-diversity and spatial β-diversity, over a period of 25 years (1990-2014) using a large dataset of 3782 stream benthic invertebrate samples collected from central European low-mountain streams. We studied a set of traits describing the ecological niche of species and their functions: body size, feeding groups, substrate preferences, flow preferences, stream zonation preferences and saprobity. Trait composition changed significantly during the study period, and we identified an overall increase in niche homogenization. Specifically, community niche specialization significantly decreased by 20.3% over the 25-year period, with declines ranging from -16.0 to -40.9% for zonation-, flow-, substrate-preferences, body size and feeding traits. Trait diversity did not change significantly, although we recorded significant decreases by -14.2% and -10.2% for flow- and substrate-preference and increases by 5.8% and 22.6% for feeding traits and zonation preference over the study period. Trait spatial β-diversity significantly decreased by -53.0%, with substrate-preference, feeding groups and flow-preference traits declining from -61.9% to -75.3% over the study period. This increased niche homogenization is likely driven by the increase of down-stream typical taxa, which are favored by warming temperatures. Further, it is in apparent contradiction with the recorded increase in abundance (+35.9%) and taxonomic richness (+39.2%) over the same period. Even such increases do not safeguard communities from undergoing niche homogenization, indicating that recovery processes may differ with regard to community taxonomic composition and traits. Our results emphasize the complexity of community responses to global change and warrant caution when founding conclusions based solely on single community metrics.

A global agenda for advancing freshwater biodiversity research

Global freshwater biodiversity is declining dramatically, and meeting the challenges of this crisis requires bold goals and the mobilisation of substantial resources. While the reasons are varied, investments in both research and conservation of freshwater biodiversity lag far behind those in the terrestrial and marine realms. Inspired by a global consultation, we identify 15 pressing priority needs, grouped into five research areas, in an effort to support informed stewardship of freshwater biodiversity. The proposed agenda aims to advance freshwater biodiversity research globally as a critical step in improving coordinated actions towards its sustainable management and conservation.

Multidecadal changes in functional diversity lag behind the recovery of taxonomic diversity

While there has been increasing interest in how taxonomic diversity is changing over time, less is known about how long-term taxonomic changes may affect ecosystem functioning and resilience. Exploring long-term patterns of functional diversity can provide key insights into the capacity of a community to carry out ecological processes and the redundancy of species' roles. We focus on a protected freshwater system located in a national park in southeast Germany. We use a high-resolution benthic macroinvertebrate dataset spanning 32 years (1983-2014) and test whether changes in functional diversity are reflected in taxonomic diversity using a multidimensional trait-based approach and regression analyses. Specifically, we asked: (i) How has functional diversity changed over time? (ii) How functionally distinct are the community's taxa? (iii) Are changes in functional diversity concurrent with taxonomic diversity? And (iv) what is the extent of community functional redundancy? Resultant from acidification mitigation, macroinvertebrate taxonomic diversity increased over the study period. Recovery of functional diversity was less pronounced, lagging behind responses of taxonomic diversity. Over multidecadal timescales, the macroinvertebrate community has become more homogenous with a high degree of functional redundancy, despite being isolated from direct anthropogenic activity. While taxonomic diversity increased over time, functional diversity has yet to catch up. These results demonstrate that anthropogenic pressures can remain a threat to biotic communities even in protected areas. The differences in taxonomic and functional recovery processes highlight the need to incorporate functional traits in assessments of biodiversity responses to global change.

Long-term data reveal unimodal responses of ground beetle abundance to precipitation and land use but no changes in taxonomic and functional diversity

While much of global biodiversity is undoubtedly under threat, the responses of ecological communities to changing climate, land use intensification, and long-term changes in both taxonomic and functional diversity over time, has still not been fully explored for many taxonomic groups, especially invertebrates. We compiled time series of ground beetles covering the past two decades from 40 sites located in five regions across Germany. We calculated site-based trends for 21 community metrics representing taxonomic and functional diversity of ground beetles, activity density (a proxy for abundance), and activity densities of functional groups. We assessed both overall and regional temporal trends and the influence of the global change drivers of temperature, precipitation, and land use on ground beetle communities. While we did not detect overall temporal changes in ground beetle taxonomic and functional diversity, taxonomic turnover changed within two regions, illustrating that community change at the local scale does not always correspond to patterns at broader spatial scales. Additionally, ground beetle activity density had a unimodal response to both annual precipitation and land use. Limited temporal change in ground beetle communities may indicate a shifting baseline, where community degradation was reached prior to the start of our observation in 1999. In addition, nonlinear responses of animal communities to environmental change present a challenge when quantifying temporal trends.

Wet grinding of invertebrate bulk samples – a scalable and cost-efficient protocol for metabarcoding and metagenomics

Most metabarcoding protocols for invertebrate bulk samples start with sample homogenisation, followed by DNA extraction, amplification of a specific marker region, and sequencing. Many of the above-mentioned laboratory steps have been verified thoroughly and best practice strategies exist, yet, no clear recommendation for the basis of almost all metabarcoding studies exists: the homogenisation of samples itself. Two different categories of devices are typically used for homogenisation: bead mills or blenders. Both have upsides and downsides. Bead mills rely on single-use plastics and therefore produce a lot of waste and are expensive. In addition to that, processing times can go up to 30 minutes making them unsuitable for large-scale studies. Blenders can handle larger sample volumes in a shorter time, and be cleaned – yet suffer from an increased risk of cross-contamination. We aimed to develop a fast, robust, cheap, and reliable sample homogenisation protocol that overcomes limitations of both approaches, i.e. does not produce difficult to discard waste and avoid single-use plastics while reducing overall costs. We tested the performance of the new protocol using six size-sorted Malaise trap samples and six unsorted stream macroinvertebrate kick-net samples. We used 14 replicates per sample and included many negative controls at different steps of the protocol to quantify the impacts of i) insufficient homogenisation and ii) cross-contamination. Our results show that 3-min homogenisation is sufficient to recover about 80% of OTUs per sample in each replicate and that a non-hazardous DIY cleaning solution provides an effective and efficient way of cleaning. The improvements of the protocol in terms of speed, ease of handling, an overall reduction of costs as well as the documented reliability and robustness make it an important candidate for sample homogenisation after sampling in particular for large-scale and regulatory metabarcoding but also metagenomics biodiversity assessments and monitoring.

Spatio-temporal niche plasticity of a freshwater invader as a harbinger of impact variability

Invasive alien fishes have detrimental ecological effects on aquatic ecosystems and the services they provide. Impacts from an invasion in a single ecosystem may differ across space and time due to variability in prey availability and environmental conditions. We hypothesize that such variability can be profound, even within a single ecosystem. Stable isotopes analysis (SIA) is commonly used to quantitatively describe the trophic niche of a species. However, spatial and temporal variability in occupied niches are often not incorporated into management strategies and policy options. Here, we used long-term monitoring data to investigate the invasion stage as well as SIA to analyse the trophic niche of the invasive channel catfish Ictalurus punctatus in Lake Kasumigaura (Japan), a long-term ecological research site (LTER), across distant sampling sites and years. We found a significant spatio-temporal variability in relative growth and isotopic niche occupation. Moreover, we defined a new index, the Isotopic Plasticity Index (IPI), which is the ratio between core and total home range of an occupied isotopic niche, to be used as a proxy for the trophic niche stretch or density. We found that this IPI varied considerably, confirming the spatio-temporal variability in trophic niches, suggesting the IPI to be an adequate new isotopic metric. Our results further provide evidence for the existence of variation across invaded landscapes, implying heterogeneous impacts on recipient native communities. Therefore, our work emphasizes the importance of exploring trophic plasticity in feeding ecology and growth as such information enables a better understanding of impacts and can inform the design and implementation of effective management responses.

Recovery from air pollution and subsequent acidification masks the effects of climate change on a freshwater macroinvertebrate community

Freshwater ecosystems are dynamic, complex systems with a multitude of physical and ecological processes and stressors which drive fluctuations on the community-level. Disentangling the effects of different processes and stressors is challenging due to their interconnected nature. However, as protected areas (i.e. national parks) are less anthropogenically impacted, they are ideal for investigating single stressors. We focus on the Bavarian Forest National Park, a Long-Term Ecological Research (LTER) site in Germany, where the major stressors are climate warming, air pollution (i.e. acidification) and bark beetle infestations. We investigated the effects of these stressors on freshwater macroinvertebrates using comprehensive long-term (1983-2014) datasets comprising high-resolution macroinvertebrate and physico-chemical data from a near-natural stream. Macroinvertebrate communities have undergone substantial changes over the past 32 years, highlighted by increases in overall community abundance (+173%) and richness (+51.6%) as well as taxonomic restructuring driven by a disproportional increase of dipterans. Prior to the year 2000, regression analyses revealed a decline in sulphate deposition and subsequent recovery from historical acidification as potential drivers of the increases in abundance and richness rather than to increases in water temperature (1.5 °C overall increase). Post 2000, however, alterations to nutrient cycling caused by bark beetle infestations coupled with warming temperatures were correlated to taxonomic restructuring and disproportional increases of dipterans at the expense of sensitive taxa such as plecopterans and trichopterans. Our results highlight the challenges when investigating the effects of climate change within a multi-stressor context. Even in conservation areas, recovery from previous disturbance might mask the effects of ongoing disturbances like climate change. Overall, we observed strong community restructuring, demonstrating that stenothermal headwater communities face additional stress due to emerging competition with tolerant taxa. Conservation efforts should consider the temporal variability of communities and their recovery from disturbances to adequately identify species vulnerable to local or widespread extinction.

Two centuries for an almost complete community turnover from native to non-native species in a riverine ecosystem

Non-native species introductions affect freshwater communities by changing community compositions, functional roles, trait occurrences and ecological niche spaces. Reconstructing such changes over long periods is difficult due to limited data availability. We collected information spanning 215 years on fish and selected macroinvertebrate groups (Mollusca and Crustacea) in the inner-Florentine stretch of the Arno River (Italy) and associated water grid, to investigate temporal changes. We identified an almost complete turnover from native to non-native fish (1800: 92% native; 2015: 94% non-native species) and macroinvertebrate species (1800: 100% native; 2015: 70% non-native species). Non-native fish species were observed ~50 years earlier compared to macroinvertebrate species, indicating phased invasion processes. In contrast, α-diversity of both communities increased significantly following a linear pattern. Separate analyses of changes in α-diversities for native and non-native species of both fish and macroinvertebrates were nonlinear. Functional richness and divergence of fish and macroinvertebrate communities decreased non-significantly, as the loss of native species was compensated by non-native species. Introductions of non-native fish and macroinvertebrate species occurred outside the niche space of native species. Native and non-native fish species exhibited greater overlap in niche space over time (62%-68%) and non-native species eventually replaced native species. Native and non-native macroinvertebrate niches overlapped to a lesser extent (15%-30%), with non-natives occupying mostly unoccupied niche space. These temporal changes in niche spaces of both biotic groups are a direct response to the observed changes in α-diversity and species turnover. These changes are potentially driven by deteriorations in hydromorphology as indicated by alterations in trait modalities. Additionally, we identified that angling played a considerable role for fish introductions. Our results support previous findings that the community turnover from native to non-native species can be facilitated by, for example, deteriorating environmental conditions and that variations in communities are multifaceted requiring more indicators than single metrics.

Combining metal and stable isotope analyses to disentangle contaminant transfer in a freshwater community dominated by alien species

Freshwater ecosystems are negatively impacted by various pollutants, from agricultural, urban and industrial wastewater, with metals being one of the largest concerns. Moreover, freshwater ecosystems are often affected by alien species introductions that can modify habitats and trophic relationships. Accordingly, the threat posed by metals interacts with those by alien species, since the latter can accumulate and transfer these substances across the food web to higher trophic levels. How metals transfer within such communities is little studied. We analysed the concentration of 14 metals/metalloids (Al, As, Cd, Co, Cr, Cu, Fe, Hg, Mg, Mn, Ni, Pb, Se, Zn, hereafter 'metal(s)') of eight fish and three crustacean species co-existing in the Arno River (Central Italy), most of which were alien. To assess the pathway of contaminants within the community, we coupled metal analysis with carbon and nitrogen stable isotope analysis derived from the same specimens. Crustaceans showed higher metal concentration than fish, except for Cd, Hg and Se that were higher in fish. We found evidence of trophic transfer for six metals (Cd, Cr, Hg, Mg, Se, Zn). Additionally, ontogenetic differences and differences among various fish tissues (muscle, liver, and gills) were found in metals concentration. Considerable biomagnification along the trophic chain was found for Hg, while other metals were found to biodilute. Using stable isotopes and Hg as a third diet tracer, we refined the estimations of consumed preys in the diet previously reconstructed with stable isotope mixing models. Alien species reach high biomass and can both survive to and accumulate high pollutants concentrations, potentially posing a risk for their predators and humans. A combined effect of environmental filtering and increased competition may potentially contribute to the disappearance of native species with lower tolerances.

Using stable isotopes to analyse extinction risks and reintroduction opportunities of native species in invaded ecosystems

Invasive non-native species have pervasive impacts on native biodiversity, including population extirpations and species extinctions. Identifying reasons why a population of a native species is extirpated following an invasion often relies on literature-based results of anecdotal observations. The well-established schemes of existing risk assessments for invasive species assume that a species' information (e.g. impacts or behavioural and biological traits) can be projected from one area to another to estimate the potential impact of a species in another environment. We used stable isotope data (δ13C, δ15N) from both invaded and uninvaded communities to predict such invasion impacts by reconstructing trophic relationships. This approach was tested on a community within a protected lake in Northern Spain where, following the introductions of non-native species, the last resident native species (the common tench Tinca tinca, the European eel Anguilla anguilla, and the whirligig beetle Gyrinus sp.) had been extirpated. Through the application of this novel approach, we found evidence that native species' declines were related to direct predation by and resource overlap with non-native species, which occurred in conjunction with habitat modification. Using this approach, we outlined the mechanisms involved in the extirpation of native species in the post-invasion period. To compensate for losses of native species induced by invasions of non-native species, native species reintroductions might be an appropriate tool. For this, we further suggested and discussed a novel approach that predicts the outcome of arising interactions by superimposing stable isotope data from alternative sources to better estimate the success of native species´ reintroductions.

SoilTemp: A global database of near-surface temperature

Current analyses and predictions of spatially explicit patterns and processes in ecology most often rely on climate data interpolated from standardized weather stations. This interpolated climate data represents long-term average thermal conditions at coarse spatial resolutions only. Hence, many climate-forcing factors that operate at fine spatiotemporal resolutions are overlooked. This is particularly important in relation to effects of observation height (e.g. vegetation, snow and soil characteristics) and in habitats varying in their exposure to radiation, moisture and wind (e.g. topography, radiative forcing or cold-air pooling). Since organisms living close to the ground relate more strongly to these microclimatic conditions than to free-air temperatures, microclimatic ground and near-surface data are needed to provide realistic forecasts of the fate of such organisms under anthropogenic climate change, as well as of the functioning of the ecosystems they live in. To fill this critical gap, we highlight a call for temperature time series submissions to SoilTemp, a geospatial database initiative compiling soil and near-surface temperature data from all over the world. Currently, this database contains time series from 7,538 temperature sensors from 51 countries across all key biomes. The database will pave the way toward an improved global understanding of microclimate and bridge the gap between the available climate data and the climate at fine spatiotemporal resolutions relevant to most organisms and ecosystem processes.

Complex and nonlinear climate-driven changes in freshwater insect communities over 42 years

The ongoing biodiversity crisis becomes evident in the widely observed decline in abundance and diversity of species, profound changes in community structure, and shifts in species' phenology. Insects are among the most affected groups, with documented decreases in abundance up to 76% in the last 25-30 years in some terrestrial ecosystems. Identifying the underlying drivers is a major obstacle as most ecosystems are affected by multiple stressors simultaneously and in situ measurements of environmental variables are often missing. In our study, we investigated a headwater stream belonging to the most common stream type in Germany located in a nature reserve with no major anthropogenic impacts except climate change. We used the most comprehensive quantitative long-term data set on aquatic insects available, which includes weekly measurements of species-level insect abundance, daily water temperature and stream discharge as well as measurements of additional physicochemical variables for a 42-year period (1969-2010). Overall, water temperature increased by 1.88 °C and discharge patterns changed significantly. These changes were accompanied by an 81.6% decline in insect abundance, but an increase in richness (+8.5%), Shannon diversity (+22.7%), evenness (+22.4%), and interannual turnover (+34%). Moreover, the community's trophic structure and phenology changed: the duration of emergence increased by 15.2 days, whereas the peak of emergence moved 13.4 days earlier. Additionally, we observed short-term fluctuations (<5 years) in almost all metrics as well as complex and nonlinear responses of the community toward climate change that would have been missed by simply using snapshot data or shorter time series. Our results indicate that climate change has already altered biotic communities severely even in protected areas, where no other interacting stressors (pollution, habitat fragmentation, etc.) are present. This is a striking example of the scientific value of comprehensive long-term data in capturing the complex responses of communities toward climate change.

Meta-analysis of multidecadal biodiversity trends in Europe

Local biodiversity trends over time are likely to be decoupled from global trends, as local processes may compensate or counteract global change. We analyze 161 long-term biological time series (15-91 years) collected across Europe, using a comprehensive dataset comprising ~6,200 marine, freshwater and terrestrial taxa. We test whether (i) local long-term biodiversity trends are consistent among biogeoregions, realms and taxonomic groups, and (ii) changes in biodiversity correlate with regional climate and local conditions. Our results reveal that local trends of abundance, richness and diversity differ among biogeoregions, realms and taxonomic groups, demonstrating that biodiversity changes at local scale are often complex and cannot be easily generalized. However, we find increases in richness and abundance with increasing temperature and naturalness as well as a clear spatial pattern in changes in community composition (i.e. temporal taxonomic turnover) in most biogeoregions of Northern and Eastern Europe.

Building Linked Spatio-Temporal Data from Vectorized Historical Maps

Historical maps provide a rich source of information for researchers in the social and natural sciences. These maps contain detailed documentation of a wide variety of natural and human-made features and their changes over time, such as the changes in the transportation networks and the decline of wetlands. It can be labor-intensive for a scientist to analyze changes across space and time in such maps, even after they have been digitized and converted to a vector format. In this paper, we present an unsupervised approach that converts vector data of geographic features extracted from multiple historical maps into linked spatio-temporal data. The resulting graphs can be easily queried and visualized to understand the changes in specific regions over time. We evaluate our technique on railroad network data extracted from USGS historical topographic maps for several regions over multiple map sheets and demonstrate how the automatically constructed linked geospatial data enables effective querying of the changes over different time periods.

Fast and Exact Rule Mining with AMIE 3

Given a knowledge base (KB), rule mining finds rules such as “If two people are married, then they live (most likely) in the same place”. Due to the exponential search space, rule mining approaches still have difficulties to scale to today’s large KBs. In this paper, we present AMIE 3, a system that employs a number of sophisticated pruning strategies and optimizations. This allows the system to mine rules on large KBs in a matter of minutes. Most importantly, we do not have to resort to approximations or sampling, but are able to compute the exact confidence and support of each rule. Our experiments on DBpedia, YAGO, and Wikidata show that AMIE 3 beats the state of the art by a factor of more than 15 in terms of runtime.

Astrea: Automatic Generation of SHACL Shapes from Ontologies

Knowledge Graphs (KGs) that publish RDF data modelled using ontologies in a wide range of domains have populated the Web. The SHACL language is a W3C recommendation that has been endowed to encode a set of either value or model data restrictions that aim at validating KG data, ensuring data quality. Developing shapes is a complex and time consuming task that is not feasible to achieve manually. This article presents two resources that aim at generating automatically SHACL shapes for a set of ontologies: (1) Astrea-KG, a KG that publishes a set of mappings that encode the equivalent conceptual restrictions among ontology constraint patterns and SHACL constraint patterns, and (2) Astrea, a tool that automatically generates SHACL shapes from a set of ontologies by executing the mappings from the Astrea-KG. These two resources are openly available at Zenodo, GitHub, and a web application. In contrast to other proposals, these resources cover a large number of SHACL restrictions producing both value and model data restrictions, whereas other proposals consider only a limited number of restrictions or focus only on value or model restrictions.

Handling Impossible Derivations During Stream Reasoning

With the rapid expansion of the Web and the advent of the Internet of Things, there is a growing need to design tools for intelligent analytics and decision making on streams of data. Logic-based frameworks like LARS allow the execution of complex reasoning on such streams, but it is paramount that the computation is completed in a timely manner before the stream expires. To reduce the runtime, we can extend the validity of inferred conclusions to the future to avoid repeated derivations, but this is not enough to avoid all sources of redundant computation. To further alleviate this problem, this paper introduces a new technique that infers the impossibility of certain derivations in the future and blocks the reasoner from performing computation that is doomed to fail anyway. An experimental analysis on microbenchmarks shows that our technique leads to a significant reduction of the reasoning runtime.

GEval: A Modular and Extensible Evaluation Framework for Graph Embedding Techniques

While RDF data are graph shaped by nature, most traditional Machine Learning (ML) algorithms expect data in a vector form. To transform graph elements to vectors, several graph embedding approaches have been proposed. Comparing these approaches is interesting for 1) developers of new embedding techniques to verify in which cases their proposal outperforms the state-of-art and 2) consumers of these techniques in choosing the best approach according to the task(s) the vectors will be used for. The comparison could be delayed (and made difficult) by the choice of tasks, the design of the evaluation, the selection of models, parameters, and needed datasets. We propose GEval, an evaluation framework to simplify the evaluation and the comparison of graph embedding techniques. The covered tasks range from ML tasks (Classification, Regression, Clustering), semantic tasks (entity relatedness, document similarity) to semantic analogies. However, GEval is designed to be (easily) extensible. In this article, we will describe the design and development of the proposed framework by detailing its overall structure, the already implemented tasks, and how to extend it. In conclusion, to demonstrate its operating approach, we consider the parameter tuning of the KGloVe algorithm as a use case.

Hyperbolic Knowledge Graph Embeddings for Knowledge Base Completion

Learning embeddings of entities and relations existing in knowledge bases allows the discovery of hidden patterns in them. In this work, we examine the contribution of geometrical space to the task of knowledge base completion. We focus on the family of translational models, whose performance has been lagging. We extend these models to the hyperbolic space so as to better reflect the topological properties of knowledge bases. We investigate the type of regularities that our model, dubbed HyperKG, can capture and show that it is a prominent candidate for effectively representing a subset of Datalog rules. We empirically show, using a variety of link prediction datasets, that hyperbolic space allows to narrow down significantly the performance gap between translational and bilinear models and effectively represent certain types of rules.

Entity Summarization with User Feedback

Semantic Web applications have benefited from entity summarization techniques which compute a compact summary for an entity by selecting a set of key triples from underlying data. A wide variety of entity summarizers have been developed. However, the quality of summaries they generate are still not satisfying, and we lack mechanisms for improving computed summaries. To address this challenge, in this paper we present the first study of entity summarization with user feedback. We consider a cooperative environment where a user reads the current entity summary and provides feedback to help an entity summarizer compute an improved summary. Our approach represents this iterative process as a Markov decision process where the entity summarizer is modeled as a reinforcement learning agent. To exploit user feedback, we represent the interdependence of triples in the current summary and the user feedback by a novel deep neural network which is incorporated into the policy of the agent. Our approach outperforms five baseline methods in extensive experiments with both real users and simulated users.

ESBM: An Entity Summarization BenchMark

Entity summarization is the problem of computing an optimal compact summary for an entity by selecting a size-constrained subset of triples from RDF data. Entity summarization supports a multiplicity of applications and has led to fruitful research. However, there is a lack of evaluation efforts that cover the broad spectrum of existing systems. One reason is a lack of benchmarks for evaluation. Some benchmarks are no longer available, while others are small and have limitations. In this paper, we create an Entity Summarization BenchMark (ESBM) which overcomes the limitations of existing benchmarks and meets standard desiderata for a benchmark. Using this largest available benchmark for evaluating general-purpose entity summarizers, we perform the most extensive experiment to date where 9 existing systems are compared. Considering that all of these systems are unsupervised, we also implement and evaluate a supervised learning based system for reference.

Equivalent Rewritings on Path Views with Binding Patterns

A view with a binding pattern is a parameterized query on a database. Such views are used, e.g., to model Web services. To answer a query on such views, the views have to be orchestrated together in execution plans. We show how queries can be rewritten into equivalent execution plans, which are guaranteed to deliver the same results as the query on all databases. We provide a correct and complete algorithm to find these plans for path views and atomic queries. Finally, we show that our method can be used to answer queries on real-world Web services.

SAShA: Semantic-Aware Shilling Attacks on Recommender Systems Exploiting Knowledge Graphs

Recommender systems (RS) play a focal position in modern user-centric online services. Among them, collaborative filtering (CF) approaches have shown leading accuracy performance compared to content-based filtering (CBF) methods. Their success is due to an effective exploitation of similarities/correlations encoded in user interaction patterns, which is computed by considering common items users rated in the past. However, their strength is also their weakness. Indeed, a malicious agent can alter recommendations by adding fake user profiles into the platform thereby altering the actual similarity values in an engineered way.

The spread of well-curated information available in knowledge graphs (\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\mathcal {KG}$$\end{document}) has opened the door to several new possibilities in compromising the security of a recommender system. In fact, \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\mathcal {KG}$$\end{document} are a wealthy source of information that can dramatically increase the attacker’s (and the defender’s) knowledge of the underlying system. In this paper, we introduce SAShA, a new attack strategy that leverages semantic features extracted from a knowledge graph in order to strengthen the efficacy of the attack to standard CF models. We performed an extensive experimental evaluation in order to investigate whether SAShA is more effective than baseline attacks against CF models by taking into account the impact of various semantic features. Experimental results on two real-world datasets show the usefulness of our strategy in favor of attacker’s capacity in attacking CF models.

Unsupervised Bootstrapping of Active Learning for Entity Resolution

Entity resolution is one of the central challenges when integrating data from large numbers of data sources. Active learning for entity resolution aims to learn high-quality matching models while minimizing the human labeling effort by selecting only the most informative record pairs for labeling. Most active learning methods proposed so far, start with an empty set of labeled record pairs and iteratively improve the prediction quality of a classification model by asking for new labels. The absence of adequate labeled data in the early active learning iterations leads to unstable models of low quality which is known as the cold start problem. In our work we solve the cold start problem using an unsupervised matching method to bootstrap active learning. We implement a thresholding heuristic that considers pre-calculated similarity scores and assigns matching labels with some degree of noise at no manual labeling cost. The noisy labels are used for initializing the active learning process and throughout the whole active learning cycle for model learning and query selection. We evaluate our pipeline with six datasets from three different entity resolution settings using active learning with a committee-based query strategy and show it successfully deals with the cold start problem. Comparing our method against two active learning baselines without bootstrapping, we show that it can additionally lead to overall improved learned models in terms of \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$F_{1}$$\end{document} score and stability.

Hybrid Reasoning Over Large Knowledge Bases Using On-The-Fly Knowledge Extraction

The success of logic-based methods for comparing entities heavily depends on the axioms that have been described for them in the Knowledge Base (KB). Due to the incompleteness of even large and well engineered KBs, such methods suffer from low recall when applied in real-world use cases. To address this, we designed a reasoning framework that combines logic-based subsumption with statistical methods for on-the-fly knowledge extraction. Statistical methods extract additional (missing) axioms for the compared entities with the goal of tackling the incompleteness of KBs and thus improving recall. Although this can be beneficial, it can also introduce noise (false positives or false negatives). Hence, our framework uses heuristics to assess whether knowledge extraction is likely to be advantageous and only activates the statistical components if this is the case. We instantiate our framework by combining lightweight logic-based reasoning implemented on top of existing triple-stores with an axiom extraction method that is based on the labels of concepts. Our work was motivated by industrial use cases over which we evaluate our instantiated framework, showing that it outperforms approaches that are only based on textual information. Besides the best combination of precision and recall, our implementation is also scalable and is currently used in an industrial production environment.

Supporting Complex Decision Making by Semantic Technologies

Complex decisions require stakeholders to identify potential decision options and collaboratively select the optimal option. Identifying potential decision options and communicating them to stakeholders is a challenging task as it requires the translation of the decision option’s technical dimension to a stakeholder-compliant language which describes the impact of the decision (e.g., financial, political). Existing knowledge-driven decision support methods generate decision options by automatically processing available data and knowledge. Ontology-based methods emerged as a sub-field in the medical domain and provide concrete instructions for given medical problems. However, the research field lacks an evaluated practical approach to support the full cycle from data and knowledge assessment to the actual decision making. This work advances the field by: (i) a problem-driven ontology engineering method which (a) supports creating the necessary ontology model for the given problem domain and (b) harmonizes relevant data and knowledge sources for automatically identifying decision options by reasoners, and (ii) an approach which translates technical decision options into a language that is understood by relevant stakeholders. Expert evaluations and real-world deployments in three different domains demonstrate the added value of this method.