Search

Smart local energy systems (SLES) have been promoted in policy as a solution to decarbonisation challenges which also bring wider benefits, such as community prosperity and energy affordability. But the combination of conditions required to enable their successful emergence and operation are still to be elaborated. This paper reports on the development of a Theory of Change (ToC) for the “societal project” of emergence of SLES with benefits. ToC is a process of making explicit the causal links by which activities lead to outcomes, surfacing assumptions, and recognising possible unintended consequences. We describe the ToC development process, involving consultation and collaboration across a research consortium. It consists of layers (e.g. users, skills, data and digital), and shows conditions considered necessary to deliver SLES, and for these to deliver wider benefits. It also provides interactive links to evidence emerging from the consortium, as well as policy/governance conditions and metrics. We reflect on potential uses of the ToC – internally to the consortium and externally – along with challenges we encountered in pursuing this approach. Policy implications relate to the importance of enabling conditions across multiple sectors, the absence of any of which could inhibit delivery of either SLES or their ensuing benefits.

Trees are long-lived organisms that integrate climate conditions across years or decades to produce secondary growth. This integration process is sometimes referred to as ‘climatic memory.’ While widely perceived, the physiological processes underlying this temporal integration, such as the storage and remobilization of non-structural carbohydrates (NSC), are rarely explicitly studied. This is perhaps most apparent when considering drought legacies (perturbed post-drought growth responses to climate), and the physiological mechanisms underlying these lagged responses to climatic extremes. Yet, drought legacies are likely to become more common if warming climate brings more frequent drought. To quantify the linkages between drought legacies, climate memory and NSC, we measured tree growth (via tree ring widths) and NSC concentrations in three dominant species across the southwestern USA. We analyzed these data with a hierarchical mixed effects model to evaluate the time-scales of influence of past climate (memory) on tree growth. We then evaluated the role of climate memory and the degree to which variation in NSC concentrations were related to forward-predicted growth during the hot 2011–2012 drought and subsequent 4-year recovery period. Populus tremuloides exhibited longer climatic memory compared to either Pinus edulis or Juniperus osteosperma, but following the 2011–2012 drought, P. tremuloides trees with relatively longer memory of temperature conditions showed larger (more negative) drought legacies. Conversely, Pinus edulis trees with longer temperature memory had smaller (less negative) drought legacies. For both species, higher NSC concentrations followed more negative (larger) drought legacies, though the relevant NSC fraction differed between P. tremuloides and P. edulis. Our results suggest that differences in tree NSC are also imprinted upon tree growth responses to climate across long time scales, which also underlie tree resilience to increasingly frequent drought events under climate change.

In trees, large uncertainties remain in how nonstructural carbohydrates (NSCs) respond to variation in water availability in natural, intact ecosystems. Variation in NSC pools reflects temporal fluctuations in supply and demand, as well as physiological coordination across tree organs in ways that differ across species and NSC fractions (e.g., soluble sugars vs starch). Using landscape-scale crown (leaves and twigs) NSC concentration measurements in three foundation tree species (Populus tremuloides, Pinus edulis, Juniperus osteosperma), we evaluated in situ, seasonal variation in NSC responses to moisture stress on three timescales: short-term (via predawn water potential), seasonal (via leaf δ13C) and annual (via current year’s ring width index). Crown NSC responses to moisture stress appeared to depend on hydraulic strategy, where J. osteosperma appears to regulate osmotic potentials (via higher sugar concentrations), P. edulis NSC responses suggest respiratory depletion and P. tremuloides responses were consistent with direct sink limitations. We also show that overly simplistic models can mask seasonal and tissue variation in NSC responses, as well as strong interactions among moisture stress at different timescales. In general, our results suggest large seasonal variation in crown NSC concentrations reflecting the multiple cofunctions of NSCs in plant tissues, including storage, growth and osmotic regulation of hydraulically vulnerable leaves. We emphasize that crown NSC pool size cannot be viewed as a simple physiological metric of stress; in situ NSC dynamics are complex, varying temporally, across species, among NSC fractions and among tissue types.

International audience; We present the WASABI 1 Song Corpus, a large corpus of songs enriched with metadata extracted from music databases on the Web, and resulting from the processing of song lyrics and from audio analysis. More specifically, given that lyrics encode an important part of the semantics of a song, we focus here on the description of the methods we proposed to extract relevant information from the lyrics, such as their structure segmentation, their topics, the explicitness of the lyrics content, the salient passages of a song and the emotions conveyed. The corpus contains 1.73M songs with lyrics (1.41M unique lyrics) annotated at different levels with the output of the above mentioned methods. The corpus labels and the provided methods can be exploited by music search engines and music professionals (e.g. journalists, radio presenters) to better handle large collections of lyrics, allowing an intelligent browsing, categorization and recommendation of songs. We demonstrate the utility and versatility of the WASABI Song Corpus in three concrete application scenarios. Together with the work on the corpus, we present the work achieved to transition the dataset into a knowledge graph, the WASABI RDF Knowledge Graph, and we show how this will enable an even richer set of applications.

Prosumer-centred electricity market models such as peer-to-peer communities can enable optimized supply and demand of locally generated electricity as well as an active participation of citizens in the energy transition. An important element of active participation is the improved ability of community members to identify and choose who they transact with in a much more granular way than is usual. Despite this key novelty and the social core of prosumer-centred markets, little is known about how citizens would trade with different actors involved in the system. Here, we report a preregistered cross-national experiment in which we investigated individual trading preferences in a peer-to-peer community, including a variety of private and non-private trading actors. Our data from the United Kingdom (n=441) and Germany (n=440) shows that set buying and selling prices strongly vary, pointing to three systematically different trading strategies that individuals apply as a function of involved trading actor. Findings moreover reveal that trading decisions are determined by individuals’ political orientation, place attachment, and climate change beliefs as well as individual differences in trust in the involved trading actor. Finally, our results illustrate high consistency in trading preferences across nations. However, nation-level differences emerged when decisions were made publicly visible, emphasising the need to consider context-effects in peer-to-peer system design. Our results have implications for the development of prosumer-centred energy models and the design of interventions to increase citizen participation across national contexts.

Starting from the first edition held in 2007, EVALITA is the initiative for the evaluation of Natural Language Processing tools for Italian. This paper describes the EVALITA4ELG project, whose main aim is at systematically collecting the resources released as benchmarks for this evaluation campaign, and making them easily accessible through the European Language Grid platform. The collection is moreover integrated with systems and baselines as a pool of web services with a common interface, deployed on a dedicated hardware infrastructure.

Song lyrics contain repeated patterns that have been proven to facilitate automated lyrics segmentation, with the final goal of detecting the building blocks (e.g., chorus, verse) of a song text. Our contribution in this article is twofold. First, we introduce a convolutional neural network (CNN)-based model that learns to segment the lyrics based on their repetitive text structure. We experiment with novel features to reveal different kinds of repetitions in the lyrics, for instance based on phonetical and syntactical properties. Second, using a novel corpus where the song text is synchronized to the audio of the song, we show that the text and audio modalities capture complementary structure of the lyrics and that combining both is beneficial for lyrics segmentation performance. For the purely text-based lyrics segmentation on a dataset of 103k lyrics, we achieve an F-score of 67.4%, improving on the state of the art (59.2% F-score). On the synchronized text–audio dataset of 4.8k songs, we show that the additional audio features improve segmentation performance to 75.3% F-score, significantly outperforming the purely text-based approaches.

Evaluation is of paramount importance in data- driven research fields such as Natural Language Processing (NLP) and Computer Vision (CV). But current evaluation practice in NLP, except for end-to-end tasks such as machine translation, spoken dialogue systems, or NLG, largely hinges on the existence of a single “ground truth” against which we can meaning- fully compare the prediction of a model. However, this assumption is flawed for two reasons. 1) In many cases, more than one answer is correct. 2) Even where there is a single answer, disagreement among annotators is ubiquitous, making it difficult to decide on a gold standard. We discuss three sources of disagreement: from the annotator, the data, and the con- text, and show how this affects even seemingly objective tasks. Current methods of adjudication, agreement, and evaluation ought to be re- considered at the light of this evidence. Some researchers now propose to address this issue by minimizing disagreement, creating cleaner datasets. We argue that such a simplification is likely to result in oversimplified models just as much as it would do for end-to-end tasks such as machine translation. Instead, we suggest that we need to improve today’s evaluation practice to better capture such disagreement. Datasets with multiple annotations are becoming more common, as are methods to integrate disagreement into modeling. The logical next step is to extend this to evaluation.

International audience; The Parental Advisory Label (PAL) is a warning label that is placed on audio recordings inrecognition of profanity or inappropriate references, with the intention of alerting parents of material potentially unsuitable for children.Since 2015, digital providers – such as iTunes,Spotify, Amazon Music and Deezer – also follow PAL guidelines and tag such tracks as “explicit”. Nowadays, such labelling is carried out mainly manually on voluntary basis, with the drawbacks of being time consuming and therefore costly, error prone and partly a subjective task. In this paper, we compare auto-mated methods ranging from dictionary-basedlookup to state-of-the-art deep neural networks to automatically detect explicit contents in English lyrics. We show that more complex models perform only slightly better on this task, and relying on a qualitative analysis of thedata, we discuss the inherent hardness and subjectivity of the task.

Given the peculiar structure of songs, applying generic text summarization methods to lyrics can lead to the generation of highly redundant and incoherent text. In this paper, we propose to enhance state-of-the-art text summarization approaches with a method inspired by audio thumbnailing. Instead of searching for the thumbnail clues in the audio of the song, we identify equivalent clues in the lyrics. We then show how these summaries that take into account the audio nature of the lyrics outperform the generic methods according to both an automatic evaluation and human judgments.

Understanding impacts of drought on tree growth and forest health is of major concern given projected climate change. Droughts may become more common in the Southwest due to extreme temperatures that will drive increased evapotranspiration and lower soil moisture, in combination with uncertain precipitation changes. Utilizing ~1.3 million tree-ring widths from the International Tree Ring Data Bank representing 10 species (eight conifers, two oaks) in the Southwest, we evaluated the effects of drought on tree growth. We categorized ring widths by formation year in relation to drought (pre-drought, drought year, and post-drought), and we used a mixed-effects model to estimate the effects of current and antecedent precipitation and temperature on tree growth during the post-drought recovery period. This allowed us to assess changes in sensitivity of tree growth to precipitation and temperature at multiple timescales following multiple droughts, and to evaluate drought resistance and recovery in these species. The effects of precipitation and temperature on ring widths following drought varied among species and time since drought. Across species, 16% of the climate effects (i.e., “sensitivities”) were significantly different from their pre-drought values. Species differed, with some showing increased sensitivities to precipitation and temperature following drought, and others showing decreased sensitivities. Furthermore, some species (e.g., Abies concolor and Pinus ponderosa) showed low resistance and slow recovery, with changes in growth sensitivities persisting up to 5 yr; others (e.g., Juniper spp.) showed high resistance, such that their climatic sensitivities did not change. Among species, the importance of different antecedent climate variables changed with time since drought. Though a majority of species responded positively to same-year precipitation pre-drought, all 10 species were positively affected by same-year precipitation the second year after drought. Our results demonstrate tree growth sensitivities vary among species and with time since drought, raising questions about physiological mechanisms and implications for forest health under future drought.

Purpose:C-reactive protein (CRP) and white blood cell (WBC) count are routine blood chemistry parameters in monitoring infection. Little is known about the natural history of their serum levels in conservative and operative spondylodiscitis treatment.Methods:Pre- and postoperative serum levels of CRP and WBC count in 145 patients with spondylodiscitis were retrospectively assessed. One hundred and four patients were treated by debridement, spondylodesis, and an antibiotic regime, 41 only with a brace and antibiotics. The results of the surgical group were compared to 156 patients fused for degenerative disc disease (DDD).Results:Surgery had a significant effect on peak postoperative CRP levels. In surgically managed patients, CRP peaked at 2–3 days after surgery (spondylodiscitis: pre-OP: 90 mg/dl vs. post-OP days 2–3: 146 mg/dl; DDD: 9 mg/dl vs. 141 mg/dl; p < 0.001), followed by a sharp decline. Although values were higher for spondylodiscitis patients, dynamics of CRP values were similar in both groups. Nonoperative treatment showed a slower decline. Surgically managed spondylodiscitis showed a higher success rate in identifying bacteria. Specific antibiotic treatment led to a more predictable decline of CRP values. WBC did not show an interpretable profile.Conclusion:CRP is a predictable serum parameter in patients with spondylodiscitis. WBC count is unspecific. Initial CRP increase after surgery is of little value in monitoring infection. A preoperative CRP value, and control once during the first 3 days after surgery is sufficient. Closer monitoring should then be continued. Should a decline not be observed, therapy needs to be scrutinized, antibiotic treatment reassessed, and concomitant infection contemplated.

This paper describes the approach of the D2KLab team to the RecSys Challenge 2018 that focuses on the task of playlist completion. We propose an ensemble strategy of different recurrent neural networks leveraging pre-trained embeddings representing tracks, artists, albums, and titles as inputs. We also use lyrics from which we extract semantic and stylistic features that we fed into the network for the creative track. The RNN learns a probabilistic model from the sequences of items in the playlist, which is then used to predict the most likely tracks to be added to the playlist. Concerning the playlists without tracks, we implemented a fall-back strategy called Title2Rec that generates recommendations using only the playlist title. We optimized the RNN, Title2Rec, and the ensemble approach on a validation set, tuning hyper-parameters such as the optimizer algorithm, the learning rate, and the generation strategy. This approach is effective in predicting tracks for a playlist and flexible to include diverse types of inputs, but it is also computationally demanding in the training phase.

The study objectives were: to investigate current cervical spine radiographic imaging practices in conscious adult patients with suspected neck injury; reasons behind variation and consideration of dose estimates were explored. Comparison with a previous survey 19 has been made. Questionnaires were sent to superintendent radiographers responsible for accident and emergency X-ray departments in English trusts with over 8500 emergency admissions per year, with a response rate of 97% (n ¼ 181/186). Departmental cervical spine imaging protocols were reported by 82% of respondents. None use fewer than the three standard projections; if the cervicothoracic junction (C7/T1), is not adequately demonstrated 87% use swimmers projections, 9% supine obliques, 3% CT alone. Following projectional radiography, 97% perform CT. A significant (p ¼ 0.018) increase was found since 1999 19 in CT use once the swimmers projection fails; fewer now use obliques at this point, continuing with CT instead. No significant difference (p ¼ 0.644) was found in choice of first supplementary radiographs; despite British Trauma Society’s 11 recommendation to undertake supine obliques, swimmers remain the most widespread technique.

The UK is home to a substantial number of heritage and tourist railways, which make a significant contribution to their local economies. They are mostly constructed on the routes of closed lines, and include large numbers of earthworks of uncertain construction and unknown strength. Recently, there have been earthwork collapses, most notably on the Gloucester and Warwickshire Railway during 2010 and 2011. The Office of Rail Regulation has also noted a number of safety incidents on heritage railways, all attributable to management failures. This paper describes an analysis of the Victorian earthworks on the Bo'ness and Kinneil Railway, a 8 km-long heritage railway in central Scotland. The analysis and risk prioritisation method used by Network Rail was found to be unsuitable for direct application to heritage railways, owing to the different operating context. A new system was therefore developed, removing some risk factors from the Network Rail approach, adding others, and modifying further ones. The new system was successfully applied, and the Bo'ness and Kinneil Railway earthworks were found to be generally stable and safe.

In our society every second polytraumatized patient is a chronic alcoholic. A patient's alcohol-related history is often unavailable and laboratory markers are not sensitive or specific enough to detect alcohol-dependent patients who are at risk of developing alcohol withdrawal syndrome (AWS) during their post-traumatic intensive care unit (ICU) stay. Previously, it has been found that plasma levels of norharman are elevated in chronic alcoholics. We investigated whether beta-carbolines, i.e. harman and norharman levels, could identify chronic alcoholics following trauma and whether possible changes during ICU stay could serve as a predictor of deterioration of clinical status. Sixty polytraumatized patients were transferred to the ICU following admission to the emergency room and subsequent surgery. Chronic alcoholics were included only if they met the DSM-III-R and ICD-10 criteria for alcohol dependence or chronic alcohol abuse/harmful use and their daily ethanol intake was > or =60 g. Harman and norharman levels were assayed on admission and on days 2, 4, 7 and 14 in the ICU. Harman and norharman levels were determined by high pressure liquid chromatography. Elevated norharman levels were found in chronic alcoholics (n = 35) on admission to the hospital and remained significantly elevated during their ICU stay. The area under the curves (AUC) showed that norharman was comparable to carbohydrate-deficient transferrin (CDT) and superior to conventional laboratory markers in detecting chronic alcoholics. Seventeen chronic alcoholics developed AWS; 16 of these patients experienced hallucinations or delirium. Norharman levels were significantly increased on days 2 and 4 in the ICU in patients who developed AWS compared with those who did not. An increase in norharman levels preceded hallucinations or delirium with a median period of approximately 3 days. The findings that elevated norharman levels are found in chronic alcoholics, that the AUC was in the range of CDT on admission and that norharman levels remained elevated during the ICU stay, support the view that norharman is a specific marker for alcoholism in traumatized patients. Since norharman levels increased prior to the onset of hallucinations and delirium it seems reasonable to investigate further the potential role of norharman as a possible substance which triggers AWS.