Search

Berth allocation problems are amongst the most important problems occurring in port terminals, and they are greatly affected by several unpredictable events. As a result, the study of these problems under uncertainty has been a target of more and more researchers. Following this research line, we consider the berth allocation problem under uncertain handling times. A distributionally robust two-stage model is presented to minimize the worst-case of the expected sum of delays with respect to a set of possible probability distributions of the handling times. The solutions of the proposed model are obtained by an exact decomposition algorithm for which several improvements are discussed. An adaptation of the proposed algorithm for the case where the assumption of relatively complete recourse fails is also presented. Extensive computational tests are reported to evaluate the effectiveness of the proposed approach and to compare the solutions obtained with those resulting from the stochastic and robust approaches. • A distributionally robust model is introduced for the berth allocation problem. • A Wasserstein distance is used to model the ambiguity set. • An exact decomposition algorithm is developed to solve the problem. • Strategies to improve the performance of the solution procedure are discussed. • Computational results show the applicability of the solution approaches. [ABSTRACT FROM AUTHOR]

Variational inference methods have been shown to lead to significant improvements in the computational efficiency of approximate Bayesian inference in mixed multinomial logit models when compared to standard Markov-chain Monte Carlo (MCMC) methods without increasing estimation bias. However, despite their demonstrated efficiency gains, existing methods still suffer from important limitations that prevent them to scale to large datasets, while providing the flexibility to allow for rich prior distributions and to capture complex posterior distributions. To effectively scale Bayesian inference in Mixed Multinomial Logit models to large datasets, we propose an Amortized Variational Inference approach that leverages stochastic backpropagation, automatic differentiation and GPU-accelerated computation. Moreover, we show how normalizing flows can be used to increase the flexibility of the variational posterior approximations. Through an extensive simulation study and real data for transport mode choice from London, we empirically show that the proposed approach is able to achieve computational speedups of multiple orders of magnitude over traditional maximum simulated likelihood estimation (MSLE) and MCMC approaches for large datasets without compromising estimation accuracy. • We propose an Amortized Variational Inference approach for Mixed Multinomial Logit. • It uses stochastic backpropagation, automatic differentiation and GPU-acceleration. • Normalizing flows can increase the flexibility of the variational approximations. • Computational speedups of orders of magnitude over MSLE and MCMC approaches. • Scales to very large datasets without compromising estimation accuracy. [ABSTRACT FROM AUTHOR]

In chronic neurological conditions, wearable/portable devices have potential as innovative tools to detect subtle early disease manifestations and disease fluctuations for the purpose of clinical diagnosis, care and therapeutic development. Huntington's disease (HD) has a unique combination of motor and non-motor features which, combined with recent and anticipated therapeutic progress, gives great potential for such devices to prove useful. The present work aims to provide a comprehensive account of the use of wearable/portable devices in HD and of what they have contributed so far. We conducted a systematic review searching MEDLINE, Embase, and IEEE Xplore. Thirty references were identified. Our results revealed large variability in the types of sensors used, study design, and the measured outcomes. Digital technologies show considerable promise for therapeutic research and clinical management of HD. However, more studies with standardized devices and harmonized protocols are needed to optimize the potential applicability of wearable/portable devices in HD. [ABSTRACT FROM AUTHOR]

Highlights • Method for precise bus travel time prediction using deep learning. • Models both cross-link (spatial) and cross-temporal correlations. • Designed for urban areas where congestion, events, etc. highly influence flow. • Empirically evaluated on large dataset from the Greater Copenhagen. • Significantly outperforms the compared baseline methods. Abstract Accurate and reliable travel time predictions in public transport networks are essential for delivering an attractive service that is able to compete with other modes of transport in urban areas. The traditional application of this information, where arrival and departure predictions are displayed on digital boards, is highly visible in the city landscape of most modern metropolises. More recently, the same information has become critical as input for smart-phone trip planners in order to alert passengers about unreachable connections, alternative route choices and prolonged travel times. More sophisticated Intelligent Transport Systems (ITS) include the predictions of connection assurance, i.e. an expert system that will decide to hold services to enable passenger exchange, in case one of the services is delayed up to a certain level. In order to operate such systems, and to ensure the confidence of passengers in the systems, the information provided must be accurate and reliable. Traditional methods have trouble with this as congestion, and thus travel time variability, increases in cities, consequently making travel time predictions in urban areas a non-trivial task. This paper presents a system for bus travel time prediction that leverages the non-static spatio-temporal correlations present in urban bus networks, allowing the discovery of complex patterns not captured by traditional methods. The underlying model is a multi-output, multi-time-step, deep neural network that uses a combination of convolutional and long short-term memory (LSTM) layers. The method is empirically evaluated and compared to other popular approaches for link travel time prediction and currently available services, including the currently deployed model at Movia, the regional public transport authority in Greater Copenhagen. We find that the proposed model significantly outperforms all the other methods we compare with, and is able to detect small irregular peaks in bus travel times very quickly. [ABSTRACT FROM AUTHOR]

Introduction: Botulinum toxin (BoNT) is the first line therapy for cervical dystonia (CD), with most patients receiving many treatment sessions, and so come to recognize and expect the benefits and harms of BoNT, making it difficult to separate which adverse events (AEs) are driven by BoNT and which come from patients' expectations.Methods: Using the results of three Cochrane systematic reviews of randomized controlled trials (RCTs) we pooled results to calculate the risk of general and specific AEs associated with BoNT, and the proportion of AEs that cannot be pharmacologically attributed to BoNT.Results: Fifteen RCTs, enrolling 1604 patients, were included. BoNT was associated with an increased risk of AEs, but 79% of this increased risk cannot be pharmacologically attributed to BoNT.Conclusions: Patients with CD attach a considerable expectation of harm due to BoNT, reflected in the large proportion of non-pharmacologically-mediated AEs. [ABSTRACT FROM AUTHOR]

Highlights • Heteroscedastic Gaussian processes (HGP) allow modeling uncertainty in traffic data. • Sample size (flow) is highly correlated with the variance in crowdsourced speeds. • Proposed SRC-HGPs explore sample size information (e.g. vehicles per minute). • SRC-HGPs are shown to produce substantially better predictive distributions. • SRC-HGPs outperform state-of-the-art for speed imputation and short-term forecasting. Abstract Accurately modeling traffic speeds is a fundamental part of efficient intelligent transportation systems. Nowadays, with the widespread deployment of GPS-enabled devices, it has become possible to crowdsource the collection of speed information to road users (e.g. through mobile applications or dedicated in-vehicle devices). Despite its rather wide spatial coverage, crowdsourced speed data also brings very important challenges, such as the highly variable measurement noise in the data due to a variety of driving behaviors and sample sizes. When not properly accounted for, this noise can severely compromise any application that relies on accurate traffic data. In this article, we propose the use of heteroscedastic Gaussian processes (HGP) to model the time-varying uncertainty in large-scale crowdsourced traffic data. Furthermore, we develop a HGP conditioned on sample size and traffic regime (SSRC-HGP), which makes use of sample size information (probe vehicles per minute) as well as previous observed speeds, in order to more accurately model the uncertainty in observed speeds. Using 6 months of crowdsourced traffic data from Copenhagen, we empirically show that the proposed heteroscedastic models produce significantly better predictive distributions when compared to current state-of-the-art methods for both speed imputation and short-term forecasting tasks. [ABSTRACT FROM AUTHOR]

Introduction: The age of diagnosis of Huntington's disease (HD) varies among individuals with the same HTT CAG-repeat expansion size. We investigated whether early-life events, like perinatal insults or neurodevelopmental disorders, influence the diagnosis age.Methods: We used data from 13,856 participants from REGISTRY and Enroll-HD, two large international multicenter observational studies. Disease-free survival analyses of mutation carriers with an HTT CAG repeat expansion size above and including 36 were computed through Kaplan-Meier estimates of median time until an HD diagnosis. Comparisons between groups were computed using a Cox proportional hazard survival model adjusted for CAG-repeat expansion length. We also assessed whether the group effect depended on gender and the affected parent.Results: Insults in the perinatal period were associated with an earlier median age of diagnosis of 45.00 years (95%CI: 42.07-47.92) compared to 51.00 years (95%CI: 50.68-51.31) in the reference group, with a CAG-adjusted hazard ratio of 1.61 (95%CI: 1.26-2.06). Neurodevelopmental disorders were also associated with an earlier median age of diagnosis than the reference group of 47.00 years (95% CI: 43.38-50.62) with a CAG-adjusted hazard ratio of 1.42 (95%CI: 1.16-1.75). These associations did not change significantly with gender or affected parent.Conclusions: These results, derived from large observational datasets, show that perinatal insults and neurodevelopmental disorders are associated with earlier ages of diagnosis of magnitudes similar to the effects of known genetic modifiers of HD. Given their clear temporal separation, these early events may be causative of earlier HD onset, but further research is needed to prove causation. [ABSTRACT FROM AUTHOR]

Electric vehicle charging demand models, with charging records as input, will inherently be biased toward the supply of available chargers. These models often fail to account for demand lost from occupied charging stations and competitors. The lost demand suggests that the actual demand is likely higher than the charging records reflect, i.e., the true demand is latent (unobserved), and the observations are censored. As a result, machine learning models that rely on these observed records for forecasting charging demand may be limited in their application in future infrastructure expansion and supply management, as they do not estimate the true demand for charging. We propose using censorship-aware models to model charging demand to address this limitation. These models incorporate censorship in their loss functions and learn the true latent demand distribution from observed charging records. We study how occupied charging stations and competing services censor demand using GPS trajectories from cars in Copenhagen, Denmark. We find that censorship occurs up to 61% of the time in some areas of the city. We use the observed charging demand from our study to estimate the true demand and find that censorship-aware models provide better prediction and uncertainty estimation of actual demand than censorship-unaware models. We suggest that future charging models based on charging records should account for censoring to expand the application areas of machine learning models in supply management and infrastructure expansion. • Observed EV charging demand is limited to the supply of charging stations. • Censored Regression models proposed to model the true demand. • Graph neural networks are extended to model the true latent demand. • Experiments on how demand is censored in Copenhagen, Denmark. • Censorship aware models offer better modelling the latent demand for charging. [ABSTRACT FROM AUTHOR]

Background: Blood biomarkers of neuronal damage could facilitate clinical management of and therapeutic development for Huntington's disease. We investigated whether neurofilament light protein NfL (also known as NF-L) in blood is a potential prognostic marker of neurodegeneration in patients with Huntington's disease.Methods: We did a retrospective analysis of healthy controls and carriers of CAG expansion mutations in HTT participating in the 3-year international TRACK-HD study. We studied associations between NfL concentrations in plasma and clinical and MRI neuroimaging findings, namely cognitive function, motor function, and brain volume (global and regional). We used random effects models to analyse cross-sectional associations at each study visit and to assess changes from baseline, with and without adjustment for age and CAG repeat count. In an independent London-based cohort of 37 participants (23 HTT mutation carriers and 14 controls), we further assessed whether concentrations of NfL in plasma correlated with those in CSF.Findings: Baseline and follow-up plasma samples were available from 97 controls and 201 individuals carrying HTT mutations. Mean concentrations of NfL in plasma at baseline were significantly higher in HTT mutation carriers than in controls (3·63 [SD 0·54] log pg/mL vs 2·68 [0·52] log pg/mL, p<0·0001) and the difference increased from one disease stage to the next. At any given timepoint, NfL concentrations in plasma correlated with clinical and MRI findings. In longitudinal analyses, baseline NfL concentration in plasma also correlated significantly with subsequent decline in cognition (symbol-digit modality test r=-0·374, p<0·0001; Stroop word reading r=-0·248, p=0·0033), total functional capacity (r=-0·289, p=0·0264), and brain atrophy (caudate r=0·178, p=0·0087; whole-brain r=0·602, p<0·0001; grey matter r=0·518, p<0·0001; white matter r=0·588, p<0·0001; and ventricular expansion r=-0·589, p<0·0001). All changes except Stroop word reading and total functional capacity remained significant after adjustment for age and CAG repeat count. In 104 individuals with premanifest Huntington's disease, NfL concentration in plasma at baseline was associated with subsequent clinical onset during the 3-year follow-up period (hazard ratio 3·29 per log pg/mL, 95% CI 1·48-7·34, p=0·0036). Concentrations of NfL in CSF and plasma were correlated in mutation carriers (r=0·868, p<0·0001).Interpretation: NfL in plasma shows promise as a potential prognostic blood biomarker of disease onset and progression in Huntington's disease.Funding: Medical Research Council, GlaxoSmithKline, CHDI Foundation, Swedish Research Council, European Research Council, Wallenberg Foundation, and Wolfson Foundation. [ABSTRACT FROM AUTHOR]

• Potential inoculating agents to improve nutrient acquisition aiding in the development of sustainable agricultural production. • All bacterial isolates obtained from natural habitat, greenhouse and tissue culture fixed nitrogen. • Endophytic bacteria efficiently promoted the growth of Cattleya walkeriana. Using root and leaf microbiomes that prevail in different Cattleya walkeriana orchid agroecosystems, i.e. , associated with the natural habitat, greenhouse, and in vitro cultivation, we performed the isolation of endophytic bacteria for bioprospecting of the mechanisms promoting plant and enzymatic growth. However, practically nothing is known about or use of its bacterial community. A total of 67 endophytic bacteria were isolated, and all showed biological nitrogen fixation capacity; 55.2% produced indole-3-acetic acid, 86.6% solubilized phosphate, and 74.6% solubilized zinc; 13.4% produced siderophores; and 71.6% had some enzymatic activity (protease, cellulase, and pectinase). The endophytes Paenibacillus taichungensis, Enterobacter sp., Rhizobium sp., Paenibacillus sp., Pseudomonas sp., and Paenibacillus pabuli were inoculated in acclimatizing seedlings obtained by micropropagation of C. walkeriana and have a potential use/role as plant growth promoters, as well as in morphological changes, nutrient uptake also resulting in increased antioxidant enzyme activity and non-enzymatic antioxidants. These results increase our understanding of the inner biome of C. walkeriana and suggest that this orchid is highly dependent on bacterial symbionts during its cycle. The isolated bacterial strains also have high potential as bioinoculant to improve nutrient acquisition and overall growth, contributing to a more sustainable. [Display omitted] [ABSTRACT FROM AUTHOR]

Copyright of Canadian Journal of Ophthalmology is the property of Elsevier B.V. and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

This research presents the architecture of a technology platform capable of integrating different types of data from building sensors and providing an interface to manage and operate facility devices, which is supported by advanced optimization algorithms. This interface is potentiated by a BIM-based interface presenting real-time data of the building. The solution, called 3i buildings - Intelligent, Interactive, and Immersive Buildings, is a tool to monitor and manage smart buildings, as well as optimize users experience, energy consumptions and environment quality. This is achieved by a grid of sensors and devices that continuously gather information (structural conditions of the building, occupancy, comfort of occupants, energy consumptions and CO2, COV's and Humidity levels, etc.), which is processed by predictive models able to learn over time. The 3D representation of the models allows managers to take advantage of the virtual environment, by augmenting the facility model and including information about the facility, making it easier and perceptible to users and owners, helping them to make better decisions. To support our research, the system will be installed in three different environments, Luz's hospital, Lisbon Aquarium and Norte Shopping, to test the solution under different conditions, objectives and users. In the first two cases the objectives are to monitor building air quality, consumptions and occupancy and in the Norte Shopping case the objectives are to monitor people flows, interact with tem and help the response in case of crisis according to the adopted emergency plan. These types of systems might help reducing energy consumptions as well as increasing comfort and satisfaction of occupants, maintaining a constant concentration of CO2 and humidity within the facility. The optimized algorithms will allow the system to learn, predicting and reacting to different conditions, giving a more reliable and smooth response to occupants needs. [ABSTRACT FROM AUTHOR]

• Mobility demand characterized by spatial and temporal variability. • Recurrent Flow Networks (RFN) formulated for spatio-temporal density estimation. • RFNs exhibit long-term predictions and fine-grained distributions on urban topologies. • Experiments with synthetic and real-world data demonstrate solution approach. Mobility-on-demand (MoD) systems represent a rapidly developing mode of transportation wherein travel requests are dynamically handled by a coordinated fleet of vehicles. Crucially, the efficiency of an MoD system highly depends on how well supply and demand distributions are aligned in spatio-temporal space (i.e., to satisfy user demand, cars have to be available in the correct place and at the desired time). To do so, we argue that predictive models should aim to explicitly disentangle between temporal and spatial variability in the evolution of urban mobility demand. However, current approaches typically ignore this distinction by either treating both sources of variability jointly, or completely ignoring their presence in the first place. In this paper, we propose recurrent flow networks 1 1 Code available at https://www.github.com/DanieleGammelli/recurrent-flow-nets (RFN), where we explore the inclusion of (i) latent random variables in the hidden state of recurrent neural networks to model temporal variability, and (ii) normalizing flows to model the spatial distribution of mobility demand. We demonstrate how predictive models explicitly disentangling between spatial and temporal variability exhibit several desirable properties, and empirically show how this enables the generation of distributions matching potentially complex urban topologies. [ABSTRACT FROM AUTHOR]

It is important to understand and forecast a typical or a particularly household daily consumption in order to design and size suitable renewable energy systems and energy storage. In this research for Short Term Load Forecasting (STLF) it has been used Artificial Neural Networks (ANN) and, despite the consumption unpredictability, it has been shown the possibility to forecast the electricity consumption of a household with certainty. The ANNs are recognized to be a potential methodology for modeling hourly and daily energy consumption and load forecasting. Input variables such as apartment area, numbers of occupants, electrical appliance consumption and Boolean inputs as hourly meter system were considered. Furthermore, the investigation carried out aims to define an ANN architecture and a training algorithm in order to achieve a robust model to be used in forecasting energy consumption in a typical household. It was observed that a feed-forward ANN and the Levenberg-Marquardt algorithm provided a good performance. For this research it was used a database with consumption records, logged in 93 real households, in Lisbon, Portugal, between February 2000 and July 2001, including both weekdays and weekend. The results show that the ANN approach provides a reliable model for forecasting household electric energy consumption and load profile. [ABSTRACT FROM AUTHOR]

Highlights: [•] Incident duration prediction models so far either ignore or deal lightly with incident record messages. [•] We apply Topic Modeling techniques to convert textual information from real-time incident reports into predictive attributes. [•] We apply Topic Modeling techniques to convert textual information from real-time incident reports into predictive attributes. [•] Results show that models with text analysis consistently outperform those without such feature, decreasing the error by 28%. [•] Same technique applies to other contexts where information is available in textual form (e.g. special events websites). [ABSTRACT FROM AUTHOR]

• Integration of machine learning and discrete choice models. • New choice model referred to as Gaussian process latent class choice model. • Derivation and implementation of an expectation-maximization algorithm. • More complex and flexible representation of unobserved heterogeneity. • The model improves prediction accuracy without weakening economic interpretability. We present a Gaussian Process – Latent Class Choice Model (GP-LCCM) to integrate a non-parametric class of probabilistic machine learning within discrete choice models (DCMs). Gaussian Processes (GPs) are kernel-based algorithms that incorporate expert knowledge by assuming priors over latent functions rather than priors over parameters, which makes them more flexible in addressing nonlinear problems. By integrating a Gaussian Process within a LCCM structure, we aim at improving discrete representations of unobserved heterogeneity. The proposed model would assign individuals probabilistically to behaviorally homogeneous clusters (latent classes) using GPs and simultaneously estimate class-specific choice models by relying on random utility models. Furthermore, we derive and implement an Expectation-Maximization (EM) algorithm to jointly estimate/infer the hyperparameters of the GP kernel function and the class-specific choice parameters by relying on a Laplace approximation and gradient-based numerical optimization methods, respectively. The model is tested on two different mode choice applications and compared against different LCCM benchmarks. Results show that GP-LCCM allows for a more complex and flexible representation of heterogeneity and improves both in-sample fit and out-of-sample predictive power. Moreover, behavioral and economic interpretability is maintained at the class-specific choice model level while local interpretation of the latent classes can still be achieved, although the non-parametric characteristic of GPs lessens the transparency of the model. [ABSTRACT FROM AUTHOR]

This study presents a Latent Class Choice Model (LCCM) with a flexible class membership component. Specifically, it formulates the latent classes using Gaussian-Bernoulli mixture models and investigates the impact of such formulation on the representation of heterogeneity in the choice process, goodness-of-fit measures and out-of-sample prediction accuracy of the choice models. Mixture models are model-based clustering techniques that have been widely used in areas such as machine learning, data mining and pattern recognition for clustering and classification problems. An Expectation-Maximization (EM) algorithm is derived for the estimation of the proposed model. Using two different case studies on travel mode choice behavior, the proposed model is compared to traditional discrete choice models on the basis of parameter estimates' signs, values of time, statistical goodness-of-fit measures, and cross-validation tests. Results show that mixture models improve the overall performance of latent class choice models by providing better out-of-sample predication accuracy in addition to better representations of heterogeneity without weakening the behavioral and economic interpretability of the choice models. • Demand model that combines unsupervised machine learning and econometric models. • New Gaussian-Bernoulli Mixture Latent Class Choice Model. • An Expectation-Maximization algorithm is derived and implemented. • More complex and flexible representation of unobserved heterogeneity. • The model improves prediction accuracy without weakening economic interpretability. [ABSTRACT FROM AUTHOR]

The association of different species of endophytic bacteria with the rhizosphere of the host plants can stimulate growth, development and acclimatization, offering a greater quantity of seedlings, in addition to reducing the cycle, providing economic return to the producer. The objective of this study was to evaluate the effect of introduction four bacterial isolates through inoculation into the root system in three banana cultivars (Prata Anã, Grande Naine and BRS Princesa) in the acclimatization phase. The evaluated treatments were: control (nutrient broth without bacteria); Bacillus cereus strain 1 (BC1); Bacillus cereus strain 2 (BC2); Bacillus thuringiensis (BT); Buttiauxella agrestis (BA). The morphological characteristics related to the development of the plants (total height and pseudostem diameter) were evaluated throughout the acclimatization period. After 90 days of transplanting and acclimatization, root length, leaf number, dry root weight, pseudostem and leaf, leaf area, internal carbon concentration, stomatal conductance, photosynthesis rate, transpiration rate, leaf temperature and chlorophyll were evaluated. The bacteria showed different results in relation to the studied cultivars. Considering the morphological and physiological characteristics observed in this study, B. thuringiensis for the cultivars Prata Anã and Grande Naine and the B. agrestis for the cultivar BRS Princesa are recommended for the process of acclimatization of banana seedlings, as they stimulated growth of the plant, increasing the dry mass, besides promoting the growth of roots. In this way, they improved the physiological aspects of the plants and reduced the period of acclimatization of the banana. [ABSTRACT FROM AUTHOR]

EN 1993-1-1 gives stability design rules for columns, beams and beam–columns up to S460, whereas EN 1993-1-12 gives additional guidance for S500 up to S700. Recent studies show that high strength steel members may be designed using improved buckling curves, where the enhanced behaviour is usually attributed to the improved material properties but mainly due to the more favourable residual stress distribution. The behaviour of unrestrained beams in HSS has not been widely studied. At present in EN 1993-1-1, the design rules for lateral-torsional buckling of beams are not dependent on the steel grade, meaning that the code does not distinguish between beams in conventional strength steel or HSS. In pursuit of an answer to the mentioned shortcomings, the present research is based on the experimental programme covering 12 full-scale tests, residual stress measurements, advanced numerical models and analytical derivations. The experiments cover different steel grades up to S690, welded and hot-rolled sections, homogeneous and hybrid (flanges in HSS and web in mild steel), double and mono-symmetric sections as well as variations in the cross-section class. This paper provides an overview of the experimental programme, discusses the results for lateral-torsional buckling of beams, and presents an advanced numerical model that was calibrated to the experimental results including the measured residual stress distribution and geometrical properties of the members. The numerical model was explored to assess various assumptions for the member imperfections, and these are further compared with code recommendations. • Full-scale experimental tests on high strength steel beams hybrid and homogeneous. • Lateral-torsional buckling of beams with doubly and monosymmetric cross-section. • Residual stresses and geometrical imperfection measurements. • Non-linear analyses with imperfections for welded I-section beams in high strength steel. • Comparison with design resistance according to Eurocode 3. [ABSTRACT FROM AUTHOR]

This paper presents two novel approaches for uncertainty estimation adapted and extended for the multi-link bus travel time problem. The uncertainty is modeled directly as part of recurrent artificial neural networks, but using two fundamentally different approaches: one based on Deep Quantile Regression and the other on Bayesian neural network. Both approaches use a recurrent neural network to predict multiple time steps into the future, but handle the time-dependent uncertainty estimation differently. We present a novel sampling technique in order to aggregate quantile estimates for link level travel time to yield the multi-link travel time distribution needed for a vehicle to travel from its current position to a specific downstream stop point or transfer site. To motivate the relevance of uncertainty-aware models in the domain, we focus on the connection protection application as a case study: An expert system to determine whether a bus driver should hold and wait for a connecting service, thus ensuring the connection, or break the connection and reduce its own delay. Our results show that the proposed quantile sampling method performs overall best for the 80%, 90% and 95% prediction intervals, both for a 15 min time horizon into the future (t + 1), but also for the 30 and 45 min time horizon (t + 2 and t + 3), with a constant, but very small underestimation of the uncertainty interval (1–4 pp.). However, we also show, that the Bayesian model still can outperform the DQR for specific cases. Lastly, we demonstrate how a simple decision support system can take advantage of our uncertainty-aware travel time models to prioritize the difference in travel time uncertainty for bus holding at strategic points, thus reducing the introduced delay for the connection protection application. • Presents two uncertainty estimation methods for multi-link bus travel time. • Presents a novel quantile sampling technique for multi-link aggregation. • Uses a connection assurance (hold and wait) application as a case study. • Finds that the proposed quantile method performs best for most prediction horizons. [ABSTRACT FROM AUTHOR]

Predicting the supply and demand of transport systems is vital for efficient traffic management, control, optimization, and planning. For example, predicting where from/to and when people intend to travel by taxi can support fleet managers in distributing resources; better predictions of traffic speeds/congestion allows for pro-active control measures or for users to better choose their paths. Making spatio-temporal predictions is known to be a hard task, but recently Graph Neural Networks (GNNs) have been widely applied on non-Euclidean spatial data. However, most GNN models require a predefined graph, and so far, researchers rely on heuristics to generate this graph for the model to use. In this paper, we use Neural Relational Inference to learn the optimal graph for the model. Our approach has several advantages: 1) a Variational Auto Encoder structure allows for the graph to be dynamically determined by the data, potentially changing through time; 2) the encoder structure allows the use of external data in the generation of the graph; 3) it is possible to place Bayesian priors on the generated graphs to encode domain knowledge. We conduct experiments on two datasets, namely the NYC Yellow Taxi and the PEMS-BAY road traffic datasets. In both datasets, we outperform benchmarks and show performance comparable to state-of-the-art. Furthermore, we do an in-depth analysis of the learned graphs, providing insights on what kinds of connections GNNs use for spatio-temporal predictions in the transport domain and how these connections can help interpretability. • Predicting the supply and demand of transportation is vital for efficient public transport. • Graph Neural Networks have shown great performance in predictions in transport systems. • Neural Relational Inference can infer graphs for GNNs to improve predictive performance. • Inferred graphs can be interpreted to better understand basis for model predictions. [ABSTRACT FROM AUTHOR]

High strength steels (HSS) are becoming more common in engineering practice due to their improved qualities. They are standardized by the specific parts of product standard EN10025 and soon they will be also codified in the execution standard EN1090. Regarding design using HSS, EN 1993 − 1 − 1 gives stability design rules for columns, beams and beam–columns up to S460, whereas EN 1993 − 1 − 12 gives additional guidance for S500 up to S700 (based mainly on numerical work available at the time). Existing studies on flexural buckling of welded H, I and box columns in steel grades S460 to S960, even though limited, show that improved curves can be used for members in high strength steel (HSS). Recently, within the European Project STROBE, evaluation of the European stability design rules was carried out covering columns, beams, and beam–columns. The research was based on experimental programme covering 20 full-scale tests, residual stress measurements, advanced numerical models, analytical derivations, and statistical evaluation. Finally, it was possible to justify new, more accurate recommendations for the buckling curve selection for HSS members. This paper provides a summary of the project conclusions regarding the stability design of steel members in high strength steel. • Stability design using high strength steel. • Reliability of design rules using Eurocode 3. • Non-linear analyses with imperfections for welded I-section columns in high strength steel. [ABSTRACT FROM AUTHOR]

Over the last few years, much online volunteered geographic information (VGI) has emerged and has been increasingly analyzed to understand places and cities, as well as human mobility and activity. However, there are concerns about the quality and usability of such VGI. In this study, we demonstrate a complete process that comprises the collection, unification, classification and validation of a type of VGI—online point-of-interest (POI) data—and develop methods to utilize such POI data to estimate disaggregated land use (i.e., employment size by category) at a very high spatial resolution (census block level) using part of the Boston metropolitan area as an example. With recent advances in activity-based land use, transportation, and environment (LUTE) models, such disaggregated land use data become important to allow LUTE models to analyze and simulate a person’s choices of work location and activity destinations and to understand policy impacts on future cities. These data can also be used as alternatives to explore economic activities at the local level, especially as government-published census-based disaggregated employment data have become less available in the recent decade. Our new approach provides opportunities for cities to estimate land use at high resolution with low cost by utilizing VGI while ensuring its quality with a certain accuracy threshold. The automatic classification of POI can also be utilized for other types of analyses on cities. [ABSTRACT FROM AUTHOR]

With an increased focus on minimising traffic externalities in metropolitan areas, a growing interest in environmentally friendly mobility systems has emerged, such as electric car-sharing systems. However, increasing demand and larger service areas often make it difficult to keep cars available where and when customers need them. This problem can be alleviated by predicting for how long cars stay vacant at given pick-up/drop-off locations. To maximise their usage, shared fleet operators relocate the cars to more desired locations. In this paper, we tackle the problem of predicting time to pick-up for shared cars in a probabilistic way by applying time-to-event modelling through survival analysis. Both statistical and machine learning approaches to survival regression are investigated. In addition, we propose the use of Gaussian copulas in order to model the correlation among vacant vehicles and to obtain more refined event-based predictions. First, an exploratory analysis is done to investigate the effect of various features on car vacancy time, which can provide significant insights into vacancy times and their influencing factors. Second, the Cox proportional hazards model (CPH), a linear survival model, is compared to DeepSurv, a neural-network-based survival model. To predict survival times, a two-step approach is formulated: in the upper level, a classification model is used to classify cars based on vacancy time duration and, in the lower level, time-to-event modelling is applied to each class using independent survival analysis models. Our empirical results using data from Copenhagen demonstrate that the DeepSurv model leads to a stronger fit compared to CPH. Moreover, we were able to verify that the proposed two-step approach can result in an improvement of over 15 % in performance compared to a standard one-step approach. Lastly, we demonstrate the benefits of survival models for relocation optimisation. • We predict shared cars' vacancy times in a probabilistic way with survival analysis. • Both statistical and machine learning methods to survival regression are applied. • We introduce the correlation among vacant vehicles through Gaussian copulas. • We formulate a two-step approach with classification of vacancy times. [ABSTRACT FROM AUTHOR]

• Observability of mobility demand is inherently limited by supply. • Censored regression applied to mobility demand to mitigate bias. • Censored Gaussian Process formulated for time-varying censorship. • Experiments with synthetic and real-world data demonstrate solution approach. • Benefit of preserving the censored information is measured. Transport demand is highly dependent on supply, especially for shared transport services where availability is often limited. As observed demand cannot be higher than available supply, historical transport data typically represents a biased, or censored , version of the true underlying demand pattern. Without explicitly accounting for this inherent distinction, predictive models of demand would necessarily represent a biased version of true demand, thus less effectively predicting the needs of service users. To counter this problem, we propose a general method for censorship-aware demand modeling, for which we derive a censored likelihood function capable of handling time-varying supply. We apply this method to the task of shared mobility demand prediction by incorporating the censored likelihood within a Gaussian Process model, which can flexibly approximate arbitrary functional forms. Experiments on artificial and real-world datasets show how taking into account the limiting effect of supply on demand is essential in the process of obtaining an unbiased predictive model of user demand behavior. [ABSTRACT FROM AUTHOR]

Background: Disease-modifying treatments are in development for Huntington's disease; crucial to their success is to identify a timepoint in a patient's life when there is a measurable biomarker of early neurodegeneration while clinical function is still intact. We aimed to identify this timepoint in a novel cohort of young adult premanifest Huntington's disease gene carriers (preHD) far from predicted clinical symptom onset.Methods: We did the Huntington's disease Young Adult Study (HD-YAS) in the UK. We recruited young adults with preHD and controls matched for age, education, and sex to ensure each group had at least 60 participants with imaging data, accounting for scan fails. Controls either had a family history of Huntington's disease but a negative genetic test, or no known family history of Huntington's disease. All participants underwent detailed neuropsychiatric and cognitive assessments, including tests from the Cambridge Neuropsychological Test Automated Battery and a battery assessing emotion, motivation, impulsivity and social cognition (EMOTICOM). Imaging (done for all participants without contraindications) included volumetric MRI, diffusion imaging, and multiparametric mapping. Biofluid markers of neuronal health were examined using blood and CSF collection. We did a cross-sectional analysis using general least-squares linear models to assess group differences and associations with age and CAG length, relating to predicted years to clinical onset. Results were corrected for multiple comparisons using the false discovery rate (FDR), with FDR <0·05 deemed a significant result.Findings: Data were obtained between Aug 2, 2017, and April 25, 2019. We recruited 64 young adults with preHD and 67 controls. Mean ages of participants were 29·0 years (SD 5·6) and 29·1 years (5·7) in the preHD and control groups, respectively. We noted no significant evidence of cognitive or psychiatric impairment in preHD participants 23·6 years (SD 5·8) from predicted onset (FDR 0·22-0·87 for cognitive measures, 0·31-0·91 for neuropsychiatric measures). The preHD cohort had slightly smaller putamen volumes (FDR=0·03), but this did not appear to be closely related to predicted years to onset (FDR=0·54). There were no group differences in other brain imaging measures (FDR >0·16). CSF neurofilament light protein (NfL), plasma NfL, and CSF YKL-40 were elevated in this far-from-onset preHD cohort compared with controls (FDR<0·0001, =0·01, and =0·03, respectively). CSF NfL elevations were more likely in individuals closer to expected clinical onset (FDR <0·0001).Interpretation: We report normal brain function yet a rise in sensitive measures of neurodegeneration in a preHD cohort approximately 24 years from predicted clinical onset. CSF NfL appears to be a more sensitive measure than plasma NfL to monitor disease progression. This preHD cohort is one of the earliest yet studied, and our findings could be used to inform decisions about when to initiate a potential future intervention to delay or prevent further neurodegeneration while function is intact.Funding: Wellcome Trust, CHDI Foundation. [ABSTRACT FROM AUTHOR]

Affective responses during exercise have been identified as a predictor of exercise adherence. However, research has been mostly limited to aerobic and resistance exercise. Considering that stretching activities are also an important component of physical fitness, this quasi-experimental study was designed to: 1) compare affective responses during and immediately after stretching exercises in apparently healthy adults, and 2) assess the consistency and repeatability of affect ratings obtained one week apart. For this purpose, we analyzed the Feeling Scale (FS) and Felt Arousal Scale (FAS) ratings using Time (during and after stretching) x Intensity (light, moderate, vigorous) x Stretched Muscle Group (quadriceps, hamstrings, glutes, latissimus dorsi, triceps) with repeated measures analysis of variance (ANCOVA) in 34 participants (21 males; aged 32.8 ± 8.6 years). The repeatability of FS and FAS ratings was assessed using two-way random-effects models, Intraclass Correlation Coefficients (ICC), and Bland-Altman plots. FS scores were higher following the stretching exercises, whereas FAS scores were lower, particularly in the vigorous intensity. In general, the inter-day repeatability for FS and FAS measurements was good across muscle groups. ICC tended to be higher at vigorous intensities. Ratings of core affect can be collected during static passive stretches using the FAS and FAS in ecologically valid settings. These results suggest that an adequate assessment of core affective responses to stretching activities should be performed during the exercises. • Timing matters when assessing affective ratings in response to stretching activities. • An affective rebound was detected in all stretching intensities. • Support was found for affect ratings and measurement procedures reliability. • Adequate assessment of affective responses must be performed during the stretches. [ABSTRACT FROM AUTHOR]

Rolling stock connections and route conflicts of trains at terminal stations are critical to train delay propagation and prediction on the railway networks. Previous studies primarily focused on delay prediction/propagation from a macro perspective, without considering the two factors. To address this problem, a hybrid neural network architecture, called TLF-net, is proposed to predict departure delays at original stations (DDOSs), considering the rolling-stock connection and potential route conflicts in the railway network. TLF-net consists of a transformer, a long short-term memory (LSTM), and a fully-connected neural network (FCNN) block, to separately address variables with different characteristics. Based on the real-world data from two terminal stations in China, the experimental results show that the consideration of the potential route conflicts from the network can considerably improve TLF-net's prediction performance over the model that only considers train interactions on a single railway line. Also, it is proven that arrival/departure routes of consecutive trains are crucial for delay prediction, while using the transformer block can efficiently reveal the route conflict severity between arrival/departure routes. The sensitivity analysis to influence factors demonstrates the significance of considering the rolling stock connection and potential route conflicts. Finally, to support real-time railway traffic management, a train arrival delay prediction model from our previous study is integrated to predict the input of TLF-net (i.e., train arrival delays), enabling the proposed model to dynamically predict the DDOSs. This dynamic updating lengthens the prediction horizon (the prediction time ahead), making it better support real-time train traffic control and management. [ABSTRACT FROM AUTHOR]

Highlights • Morphologic markers allowed to determine quality, maturity and fruit harvest season. • Increase in pulp weight and yield, pH, and soluble solids. • Decrease in weight and thickness, firmness and moisture. • Up to 42 days the fruits possess adequate internal quality to be consumed. Abstract This study aimed to evaluate the evolution of physical and physicochemical characteristics of white-fleshed dragon fruit during its development. Correlations between external color and the main pulp quality variables were analyzed. Important characteristics were observed throughout fruit development, such as increase in pulp weight and yield, pH, and soluble solids and decrease in weight and thickness, firmness, and moisture. The visual changes in different development stages showed their potential use as morphologic markers in the determination of fruit ripening, especially the appearance of fruit scales. Based on the analyzed variables, the ideal harvest time indicated for commercialization in the short or medium term is between 34–38 days after anthesis. After 40 days, the fruit becomes less visually attractive because of scale wilting and loss of the intense pink color of the skin; although, up to 42 days, they possess adequate internal quality to be consumed, and may be also used for processing. [ABSTRACT FROM AUTHOR]

• A multi-output neural network regression model to predict injury severity. • Natural language processing and topic modelling to identify Latent Dirichlet Allocation topics from self-reports. • Cyclists' self-reports are important for assessment of perceived safety. • Cyclist crash data from emergency forces include valurable safety information. • Perceived safety and road characteristics influence injury severity for cyclists. The use of machine learning techniques in safety research has increased as has the interest in using new data sources. This study's unique contribution is the application of text mining—focusing on perceived cyclist safety and crash occurrence in an urban environment. We analysed crash data collected by the emergency forces in the Capital Region of Denmark from 2013 to 2017 and self-reported textual data provided by cyclists from 2018 to 2019. The analysis included natural language processing and topic modelling to identify Latent Dirichlet Allocation (LDA) topics from self-reports, representing environment characteristics that cyclists' perceive as unsafe. A multi-output neural network regression model is applied to predict the injury-severity distribution of cyclists involved in crashes (measured by emergency response level [ERL]) based on the obtained topic distributions together with additional variables like cycle flow. We identified six LDA topics which address buses and cycle paths, conflicts with parked cars, roundabouts and inadequate maintenance, fast-moving cars and lack of cycle path, school zones and heavy traffic, and intersections and interactions with vehicles. Cycle flow was found to be the highest impacter on ERL prediction. However, other factors also impacted ERLs, especially school zones and heavy traffic. The results bring new insights into safety perception and actual safety for cyclists. The results contribute to a novel procedure for the joint correlation analysis using machine learning techniques on self-reported textual data thereby providing a better tool for infrastructure planning. The findings show the importance of including perceived safety in crash modelling and that authorities should focus on safety around schools and in intersections in order to improve safety for cyclists in a urban environment. [ABSTRACT FROM AUTHOR]

Multiple myeloma (MM) is a progressive plasma cell malignancy, with a range of clinical features including bone lesions, renal insufficiency, anaemia, and hypercalcaemia. Novel agents have significantly improved patient survival, however most patients will suffer multiple relapses. Although clinical challenges and economic costs of relapse are recognised, the psychological impact of relapse is not fully appreciated. Additionally, there is little information on how physicians perceive the impact of relapse on their patients’ emotional state and how this might affect patient management. Through face-to-face interviews with 50 relapsed and/or refractory MM patients and 30 haematologists across ten countries, we have used real-world evidence to explore and characterise the burden of living with MM, particularly the impact of relapsed disease. This exploratory study illustrates the impact of the disease on friends and family, and the physical and emotional burden experienced by the patient resulting from both MM and its treatment. Haematologists feel poorly equipped to deal with the emotional aspects of patient relapse, lacking the time and resources to adequately deal with these issues. Focused educational and support tools/resources targeted at both physicians and patients are required to facilitate physician–patient communication to help reduce the emotional burden of living with MM. [ABSTRACT FROM AUTHOR]

• Censored data as defining characteristic of numerous domains in science. • Heteroscedastic Multi-Output Gaussian Process formulated for censored regression. • Generalization of arbitrary likelihood functions enabled by devising a variational bound to the marginal log-likelihood. • Experiments with synthetic and real-world data demonstrate solution approach. When modelling censored observations (i.e. data in which the value of a measurement or observation is un-observable beyond a given threshold), a typical approach in current regression methods is to use a censored-Gaussian (i.e. Tobit) model to describe the conditional output distribution. In this paper, as in the case of missing data, we argue that exploiting correlations between multiple outputs can enable models to better address the bias introduced by censored data. To do so, we introduce a heteroscedastic multi-output Gaussian process model which combines the non-parametric flexibility of GPs with the ability to leverage information from correlated outputs under input-dependent noise conditions. To address the resulting inference intractability, we further devise a variational bound to the marginal log-likelihood suitable for stochastic optimization. We empirically evaluate our model against other generative models for censored data on both synthetic and real world tasks and further show how it can be generalized to deal with arbitrary likelihood functions. Results show how the added flexibility allows our model to better estimate the underlying non-censored (i.e. true) process under potentially complex censoring dynamics. [ABSTRACT FROM AUTHOR]

• Route conflicts at multi-line stations are considered from a microscopic train operation view. • Arrival/departure route data is treated as textual information and processed by word embedding technique. • The consideration of the route conflicts is proven to substantially improve the model performance. • The model exhibits higher predictive accuracy than existing models and robust performance at different multi-line stations. Multi-line stations (MLSs) are the intersections of different railway lines; they are crucial for delay propagation in railway networks. Therefore, the precise prediction of train arrival delays at the MLSs can efficiently support train operation rescheduling plans and reduce delay propagation in the railway network. The arrival routes of trains at the MLSs are critical factors for managing train arrival delays, since there may be latent route conflicts with forward arrival/departure trains. However, route conflicts will not occur at single-line stations (SLSs) that are traversed by only one railway line. Existing train delay prediction studies have considered the ways that trains arrive at/depart from stations as black boxes, but have not considered the latent route conflicts from a microscopic view. This study considers the arrival routes of predicted trains and route conflicts with forward trains, for contemplating the gap (not considering the route conflicts from other railway lines) in the existing studies. The influencing factors are separated into three categories according to the data attributes, namely, route-related variables, delay-related variables, and environment-related variables. Then, an architecture called LLCF-net is proposed, with a one-dimensional convolutional neural network (CNN) block for route-related variables, two long short-term memory (LSTM) networks for delay-related variables, and a fully connected neural network (FCNN) block for environment-related variables. Compared with the methods in exiting studies, this architecture showed the best performance for both two MLSs—GuangzhouSouth(GZS) and ChangshaSouth (CSS)—on the Chinese high-speed railway network, regardless of the consideration of route-related variables. In addition, LLCF-net is proven to have a strong predictive effectiveness and a robust performance for different delay lengths. [ABSTRACT FROM AUTHOR]

Bike-sharing systems are a rapidly developing mode of transportation and provide an efficient alternative to passive, motorized personal mobility. The asymmetric nature of bike demand causes the need for rebalancing bike stations, which is typically done during nighttime. To determine the optimal starting inventory level of a station for a given day, a User Dissatisfaction Function (UDF) models user pickups and returns as non-homogeneous Poisson processes with piece-wise linear rates. In this paper, we devise a deep generative model directly applicable in the UDF by introducing a variational Poisson recurrent neural network model (VP-RNN) to forecast future pickup and return rates. We empirically evaluate our approach against both traditional and learning-based forecasting methods on real trip travel data from the city of New York, USA, and show how our model outperforms benchmarks in terms of system efficiency and demand satisfaction. By explicitly focusing on the combination of decision-making algorithms with learning-based forecasting methods, we highlight a number of shortcomings in literature. Crucially, we show how more accurate predictions do not necessarily translate into better inventory decisions. By providing insights into the interplay between forecasts, model assumptions, and decisions, we point out that forecasts and decision models should be carefully evaluated and harmonized to optimally control shared mobility systems. • Predictive and prescriptive performance does not always align. • Predictive models should be evaluated on inventory decision quality. • Variational Poisson RNN formulated for pickup and return rate prediction. • Experiments with synthetic and real-world data demonstrate solution approach. • The misalignment between prediction and decision objectives is quantified. [ABSTRACT FROM AUTHOR]