The impact of altering restaurant and menu option position on food selected from an experimental food delivery platform: a randomised controlled trial.

BACKGROUND: Overconsumption is one of the most serious public health challenges in the UK and has been linked to increased consumption of food ordered through delivery platforms. This study tested whether repositioning foods and/or restaurant options in a simulated food delivery platform could help to reduce the energy content of users' shopping basket. METHODS: UK adult food delivery platform users (N = 9,003) selected a meal in a simulated platform. Participants were randomly allocated to a control condition (choices listed randomly) or to one of four intervention groups, (1) food options listed in ascending order of energy content, (2) restaurant options listed in ascending order of average energy content per main meal, (3) interventions 1 and 2 combined (4) interventions 1 and 2 combined, but food and restaurant options repositioned based on a kcal/price index to display options lower in energy but higher in price at the top. Gamma regressions assessed the impact of interventions on total energy content of baskets at checkout. RESULTS: The energy content of participants' baskets in the control condition was 1382 kcals. All interventions significantly reduced energy content of baskets: Compared to control, repositioning both foods and restaurants purely based on energy content of options resulted in the greatest effect (-209kcal; 95%CIs: -248,-168), followed by repositioning restaurants (-161kcal; 95%CIs: -201,-121), repositioning restaurants and foods based on a kcal/price index (-117kcals; 95%CI: -158,-74) and repositioning foods based on energy content (-88kcals; 95%CI: -130,-45). All interventions reduced the basket price compared to the control, except for the intervention repositioning restaurants and foods based on a kcal/price index, which increased the basket price. CONCLUSIONS: This proof-of-concept study suggests repositioning lower-energy options more prominently may encourage lower energy food choices in online delivery platforms and can be implemented in a sustainable business model.

Individual risk prediction of high grade prostate cancer based on the combination between total prostate-specific antigen (PSA) and free to total PSA ratio.

OBJECTIVES: Clinical practice guidelines endorse the stratification of prostate cancer (PCa) risk according to individual total prostate-specific antigen (tPSA) values and age to enhance the individual risk-benefit ratio. We defined two nomograms to predict the individual risk of high and low grade PCa by combining the assay of tPSA and %free/tPSA (%f/tPSA) in patients with a pre-biopsy tPSA between 2 and 10 µg/L. METHODS: The study cohort consisted of 662 patients that had fPSA, tPSA, and a biopsy performed (41.3% with a final diagnosis of PCa). Logistic regression including age, tPSA and %f/tPSA was used to model the probability of having high or low grade cancer by defining 3 outcome levels: no PCa, low grade (International Society of Urological Pathology grade, ISUP<3) and high grade PCa (ISUP≥3). RESULTS: The nomogram identifying patients with: (a) high vs. those with low grade PCa and without the disease showed a good discriminating capability (∼80%), but the calibration showed a risk of underestimation for predictive probabilities >30% (a considerable critical threshold of risk), (b) ISUP<3 vs. those without the disease showed a discriminating capability of 63% and overestimates predictive probabilities >50%. In ISUP 5 a possible loss of PSA immunoreactivity has been observed. CONCLUSIONS: The estimated risk of high or low grade PCa by the nomograms may be of aid in the decision-making process, in particular in the case of critical comorbidities and when the digital rectal examinations are inconclusive. The improved characterization of the risk of ISUP≥3 might enhance the use for magnetic resonance imaging in this setting.

Ensemble Conformalized Quantile Regression for Probabilistic Time Series Forecasting.

This article presents a novel probabilistic forecasting method called ensemble conformalized quantile regression (EnCQR). EnCQR constructs distribution-free and approximately marginally valid prediction intervals (PIs), which are suitable for nonstationary and heteroscedastic time series data. EnCQR can be applied on top of a generic forecasting model, including deep learning architectures. EnCQR exploits a bootstrap ensemble estimator, which enables the use of conformal predictors for time series by removing the requirement of data exchangeability. The ensemble learners are implemented as generic machine learning algorithms performing quantile regression (QR), which allow the length of the PIs to adapt to local variability in the data. In the experiments, we predict time series characterized by a different amount of heteroscedasticity. The results demonstrate that EnCQR outperforms models based only on QR or conformal prediction (CP), and it provides sharper, more informative, and valid PIs.

Age prediction by deep learning applied to Greenland halibut (Reinhardtius hippoglossoides) otolith images.

Otoliths (ear-stones) in the inner ears of vertebrates containing visible year zones are used extensively to determine fish age. Analysis of otoliths is a time-consuming and difficult task that requires the education of human experts. Human age estimates are inconsistent, as several readings by the same human expert might result in different ages assigned to the same otolith, in addition to an inherent bias between readers. To improve efficiency and resolve inconsistent results in the age reading from otolith images by human experts, an automated procedure based on convolutional neural networks (CNNs), a class of deep learning models suitable for image processing, is investigated. We applied CNNs that perform image regression to estimate the age of Greenland halibut (Reinhardtius hippoglossoides) with good results for individual ages as well as the overall age distribution, with an average CV of about 10% relative to the read ages by experts. In addition, the density distribution of predicted ages resembles the density distribution of the ground truth. By using k*l-fold cross-validation, we test all available samples, and we show that the results are rather sensitive to the choice of test set. Finally, we apply explanation techniques to analyze the decision process of deep learning models. In particular, we produce heatmaps indicating which input features that are the most important in the computation of predicted age.

Understanding Pooling in Graph Neural Networks.

Many recent works in the field of graph machine learning have introduced pooling operators to reduce the size of graphs. In this article, we present an operational framework to unify this vast and diverse literature by describing pooling operators as the combination of three functions: selection, reduction, and connection (SRC). We then introduce a taxonomy of pooling operators, based on some of their key characteristics and implementation differences under the SRC framework. Finally, we propose three criteria to evaluate the performance of pooling operators and use them to investigate the behavior of different operators on a variety of tasks.

Code-Aligned Autoencoders for Unsupervised Change Detection in Multimodal Remote Sensing Images.

Image translation with convolutional autoencoders has recently been used as an approach to multimodal change detection (CD) in bitemporal satellite images. A main challenge is the alignment of the code spaces by reducing the contribution of change pixels to the learning of the translation function. Many existing approaches train the networks by exploiting supervised information of the change areas, which, however, is not always available. We propose to extract relational pixel information captured by domain-specific affinity matrices at the input and use this to enforce alignment of the code spaces and reduce the impact of change pixels on the learning objective. A change prior is derived in an unsupervised fashion from pixel pair affinities that are comparable across domains. To achieve code space alignment, we enforce pixels with similar affinity relations in the input domains to be correlated also in code space. We demonstrate the utility of this procedure in combination with cycle consistency. The proposed approach is compared with the state-of-the-art machine learning and deep learning algorithms. Experiments conducted on four real and representative datasets show the effectiveness of our methodology.

Comparison of Three Dietary Assessment Methods to Estimate Meat Intake as Part of a Meat Reduction Intervention among Adults in the UK.

Food diaries are used to estimate meat intake at an individual level but it is unclear whether simpler methods would provide similar results. This study assessed the agreement between 7 day food diaries in which composite dishes were disaggregated to assess meat content (reference method), and two simpler methods: (1) frequency meal counts from 7 day food diaries; and (2) 7 day dietary recalls, each using standard estimated portion sizes. We compared data from a randomized controlled trial testing a meat reduction intervention. We used Bland-Altman plots to assess the level of agreement between methods at baseline and linear mixed-effects models to compare estimates of intervention effectiveness. At baseline, participants consumed 132 g/d (±75) of total meat; frequency meal counts and dietary recalls underestimated this by an average of 30 and 34 g/day, respectively. This was partially explained by an underestimation of the assumed portion size. The two simpler methods also underestimated the effect of the intervention, relative to control, though the significant effect of the intervention was unchanged. Simpler methods underestimated absolute meat intake but may be suitable for use in studies to measure the change in meat intake in individuals over time.

Replacing meat with alternative plant-based products (RE-MAP): a randomized controlled trial of a multicomponent behavioral intervention to reduce meat consumption.

BACKGROUND: Reducing meat consumption could protect the environment and human health. OBJECTIVES: We tested the impact of a behavioral intervention to reduce meat consumption. METHODS: Adult volunteers who regularly consumed meat were recruited from the general public and randomized 1:1 to an intervention or control condition. The intervention comprised free meat substitutes for 4 weeks, information about the benefits of eating less meat, success stories, and recipes. The control group received no intervention or advice on dietary change. The primary outcome was daily meat consumption after 4 weeks, assessed by a 7-day food diary, and repeated after 8 weeks as a secondary outcome. Other secondary and exploratory outcomes included the consumption of meat substitutes, cardiovascular risk factors, psychosocial variables related to meat consumption, and the nutritional composition of the diet. We also estimated the intervention's environmental impact. We evaluated the intervention using generalized linear mixed-effects models. RESULTS: Between June 2018 and October 2019, 115 participants were randomized. The baseline meat consumption values were 134 g/d in the control group and 130 g/d in the intervention group. Relative to the control condition, the intervention reduced meat consumption at 4 weeks by 63 g/d (95% CI: 44-82; P < 0.0001; n = 114) and at 8 weeks by 39 g/d (95% CI: 16-62; P = 0.0009; n = 113), adjusting for sex and baseline consumption. The intervention significantly increased the consumption of meat substitutes without changing the intakes of other principal food groups. The intervention increased intentions, positive attitudes, perceived control, and subjective norms of eating a low-meat diet and using meat substitutes, and decreased attachment to meat. At 8 weeks, 55% of intervention recipients identified as meat eaters, compared to 89% of participants in the control group. CONCLUSIONS: A behavioral program involving free meat substitutes can reduce meat intake and change psychosocial constructs consistent with a sustained reduction in meat intake.

Hierarchical Representation Learning in Graph Neural Networks With Node Decimation Pooling.

In graph neural networks (GNNs), pooling operators compute local summaries of input graphs to capture their global properties, and they are fundamental for building deep GNNs that learn hierarchical representations. In this work, we propose the Node Decimation Pooling (NDP), a pooling operator for GNNs that generates coarser graphs while preserving the overall graph topology. During training, the GNN learns new node representations and fits them to a pyramid of coarsened graphs, which is computed offline in a preprocessing stage. NDP consists of three steps. First, a node decimation procedure selects the nodes belonging to one side of the partition identified by a spectral algorithm that approximates the MAXCUT solution. Afterward, the selected nodes are connected with Kron reduction to form the coarsened graph. Finally, since the resulting graph is very dense, we apply a sparsification procedure that prunes the adjacency matrix of the coarsened graph to reduce the computational cost in the GNN. Notably, we show that it is possible to remove many edges without significantly altering the graph structure. Experimental results show that NDP is more efficient compared to state-of-the-art graph pooling operators while reaching, at the same time, competitive performance on a significant variety of graph classification tasks.

Graph Neural Networks With Convolutional ARMA Filters.

Popular graph neural networks implement convolution operations on graphs based on polynomial spectral filters. In this paper, we propose a novel graph convolutional layer inspired by the auto-regressive moving average (ARMA) filter that, compared to polynomial ones, provides a more flexible frequency response, is more robust to noise, and better captures the global graph structure. We propose a graph neural network implementation of the ARMA filter with a recursive and distributed formulation, obtaining a convolutional layer that is efficient to train, localized in the node space, and can be transferred to new graphs at test time. We perform a spectral analysis to study the filtering effect of the proposed ARMA layer and report experiments on four downstream tasks: semi-supervised node classification, graph signal classification, graph classification, and graph regression. Results show that the proposed ARMA layer brings significant improvements over graph neural networks based on polynomial filters.

The Impact of Environmental Sustainability Labels on Willingness-to-Pay for Foods: A Systematic Review and Meta-Analysis of Discrete Choice Experiments.

Food production is a major contributor to environmental damage. More environmentally sustainable foods could incur higher costs for consumers. In this review, we explore whether consumers are willing to pay (WTP) more for foods with environmental sustainability labels ('ecolabels'). Six electronic databases were searched for experiments on consumers' willingness to pay for ecolabelled food. Monetary values were converted to Purchasing Power Parity dollars and adjusted for country-specific inflation. Studies were meta-analysed and effect sizes with confidence intervals were calculated for the whole sample and for pre-specified subgroups defined as meat-dairy, seafood, and fruits-vegetables-nuts. Meta-regressions tested the role of label attributes and demographic characteristics on participants' WTP. Forty-three discrete choice experiments (DCEs) with 41,777 participants were eligible for inclusion. Thirty-five DCEs (n = 35,725) had usable data for the meta-analysis. Participants were willing to pay a premium of 3.79 PPP$/kg (95%CI 2.7, 4.89, p ≤ 0.001) for ecolabelled foods. WTP was higher for organic labels compared to other labels. Women and people with lower levels of education expressed higher WTP. Ecolabels may increase consumers' willingness to pay more for environmentally sustainable products and could be part of a strategy to encourage a transition to more sustainable diets.

The Effects of Environmental Sustainability Labels on Selection, Purchase, and Consumption of Food and Drink Products: A Systematic Review.

This review assessed the effects of environmental labels on consumers' demand for more sustainable food products. Six electronic databases were searched for experimental studies of ecolabels and food choices. We followed standard Cochrane methods and results were synthesized using vote counting. Fifty-six studies (N = 42,768 participants, 76 interventions) were included. Outcomes comprised selection (n = 14), purchase (n = 40) and consumption (n = 2). The ecolabel was presented as text (n = 36), logo (n = 13) or combination (n = 27). Message types included: organic (n = 25), environmentally sustainable (n = 27), greenhouse gas emissions (n = 17), and assorted "other" message types (n = 7). Ecolabels were tested in actual (n = 15) and hypothetical (n = 41) environments. Thirty-nine studies received an unclear or high RoB rating. Sixty comparisons favored the intervention and 16 favored control. Ecolabeling with a variety of messages and formats was associated with the selection and purchase of more sustainable food products.

Definition of Outcome-Based Prostate-Specific Antigen (PSA) Thresholds for Advanced Prostate Cancer Risk Prediction.

We defined prostate-specific antigen (PSA) thresholds from a well calibrated risk prediction model for identifying and excluding advanced prostate cancer (PCa). We retrieved 902 biopsied patients with a pre-biopsy PSA determination (Roche assay). A logistic regression model predictive for PCa including the main effects [i.e., PSA, age, histological evidence of glandular inflammation (GI)] was built after testing the accuracy by calibration plots and Hosmer-Lemeshow test for goodness of fit. PSA thresholds were derived by assuming a diagnostic sensitivity of 95% (rule-out) and 80% (rule-in) for overall and advanced/poorly differentiated PCa. In patients without GI, serum PSA concentrations ≤ 4.1 (<65 years old) and ≤3.7 µg/L (≥65 years old) excluded an advanced PCa (defined as Gleason score ≥ 7 at biopsy), with a negative predictive value of 95.1% [95% confidence interval (CI): 83.0-98.7] and 88.8% (CI: 80.2-93.9), respectively, while PSA > 5.7 (<65) and >6.1 µg/L (≥65) should address biopsy referral. In presence of GI, PSA did not provide a valid estimate for risk of advanced cancer because of its higher variability and the low pre-test probability of PCa. The proposed PSA thresholds may support biopsy decision except for patients with asymptomatic prostatitis who cannot be pre-biopsy identified.

Estimation of Excess Mortality and Years of Life Lost to COVID-19 in Norway and Sweden between March and November 2020.

We estimate the weekly excess all-cause mortality in Norway and Sweden, the years of life lost (YLL) attributed to COVID-19 in Sweden, and the significance of mortality displacement. We computed the expected mortality by taking into account the declining trend and the seasonality in mortality in the two countries over the past 20 years. From the excess mortality in Sweden in 2019/20, we estimated the YLL attributed to COVID-19 using the life expectancy in different age groups. We adjusted this estimate for possible displacement using an auto-regressive model for the year-to-year variations in excess mortality. We found that excess all-cause mortality over the epidemic year, July 2019 to July 2020, was 517 (95%CI = (12, 1074)) in Norway and 4329 [3331, 5325] in Sweden. There were 255 COVID-19 related deaths reported in Norway, and 5741 in Sweden, that year. During the epidemic period of 11 March-11 November, there were 6247 reported COVID-19 deaths and 5517 (4701, 6330) excess deaths in Sweden. We estimated that the number of YLL attributed to COVID-19 in Sweden was 45,850 [13,915, 80,276] without adjusting for mortality displacement and 43,073 (12,160, 85,451) after adjusting for the displacement accounted for by the auto-regressive model. In conclusion, we find good agreement between officially recorded COVID-19 related deaths and all-cause excess deaths in both countries during the first epidemic wave and no significant mortality displacement that can explain those deaths.

Reservoir Computing Approaches for Representation and Classification of Multivariate Time Series.

Classification of multivariate time series (MTS) has been tackled with a large variety of methodologies and applied to a wide range of scenarios. Reservoir computing (RC) provides efficient tools to generate a vectorial, fixed-size representation of the MTS that can be further processed by standard classifiers. Despite their unrivaled training speed, MTS classifiers based on a standard RC architecture fail to achieve the same accuracy of fully trainable neural networks. In this article, we introduce the reservoir model space, an unsupervised approach based on RC to learn vectorial representations of MTS. Each MTS is encoded within the parameters of a linear model trained to predict a low-dimensional embedding of the reservoir dynamics. Compared with other RC methods, our model space yields better representations and attains comparable computational performance due to an intermediate dimensionality reduction procedure. As a second contribution, we propose a modular RC framework for MTS classification, with an associated open-source Python library. The framework provides different modules to seamlessly implement advanced RC architectures. The architectures are compared with other MTS classifiers, including deep learning models and time series kernels. Results obtained on the benchmark and real-world MTS data sets show that RC classifiers are dramatically faster and, when implemented using our proposed representation, also achieve superior classification accuracy.

Intervention Fatigue is the Primary Cause of Strong Secondary Waves in the COVID-19 Pandemic.

As of November 2020, the number of COVID-19 cases was increasing rapidly in many countries. In Europe, the virus spread slowed considerably in the late spring due to strict lockdown, but a second wave of the pandemic grew throughout the fall. In this study, we first reconstruct the time evolution of the effective reproduction numbers R(t) for each country by integrating the equations of the classic Susceptible-Infectious-Recovered (SIR) model. We cluster countries based on the estimated R(t) through a suitable time series dissimilarity. The clustering result suggests that simple dynamical mechanisms determine how countries respond to changes in COVID-19 case counts. Inspired by these results, we extend the simple SIR model for disease spread to include a social response to explain the number X(t) of new confirmed daily cases. In particular, we characterize the social response with a first-order model that depends on three parameters ν1,ν2,ν3. The parameter ν1 describes the effect of relaxed intervention when the incidence rate is low; ν2 models the impact of interventions when incidence rate is high; ν3 represents the fatigue, i.e., the weakening of interventions as time passes. The proposed model reproduces typical evolving patterns of COVID-19 epidemic waves observed in many countries. Estimating the parameters ν1,ν2,ν3 and initial conditions, such as R0, for different countries helps to identify important dynamics in their social responses. One conclusion is that the leading cause of the strong second wave in Europe in the fall of 2020 was not the relaxation of interventions during the summer, but rather the failure to enforce interventions in the fall.

Collapse of Orthotropic Spherical Shells.

We report on the buckling and subsequent collapse of orthotropic elastic spherical shells under volume and pressure control. Going far beyond what is known for isotropic shells, a rich morphological phase space with three distinct regimes emerges upon variation of shell slenderness and degree of orthotropy. Our extensive numerical simulations are in agreement with experiments using fabricated polymer shells. The shell buckling pathways and corresponding strain energy evolution are shown to depend strongly on material orthotropy. We find surprisingly robust orthotropic structures with strong similarities to stomatocytes and tricolpate pollen grains, suggesting that the shape of several of Nature's collapsed shells could be understood from the viewpoint of material orthotropy.

Replacing meat with alternative plant-based products (RE-MAPs): protocol for a randomised controlled trial of a behavioural intervention to reduce meat consumption.

INTRODUCTION: Reducing meat consumption could contribute towards preventing some chronic conditions and protecting the natural environment. This study will examine the effectiveness of a behavioural intervention to reduce meat consumption. METHODS AND ANALYSES: Replacing meat with alternative plant-based product is a randomised controlled trial comparing a behavioural intervention to reduce meat consumption with a no intervention control condition. Eligible volunteers will be recruited from the general public through advertisement and randomised in a 1:1 ratio to receive no intervention or a 4-week intervention comprising the provision of free plant-based meat alternatives, written information on the health and environmental benefits of eating less meat, success stories of people who reduced their meat consumption and recipes. The primary outcome is the change in meat consumption at 4 weeks (T1) from baseline. Secondary and exploratory outcomes include changes in meat consumption at 8 weeks (T2) from baseline and changes from the baseline to both follow-up in other aspects of participants diet, putative psychosocial determinants of eating a low meat diet and of using meat substitutes and biomarkers of health risk, including blood lipid profiles, blood pressure, weight and body composition. Linear models will be employed to explore whether the changes in each of the aforementioned outcomes differ significantly between the control and intervention group. Qualitative interviews on a subsample of participants receiving the intervention will evaluate their experiences of the intervention and help to identify the mechanisms through which the intervention reduced meat consumption or the barriers preventing the intervention to aid this dietary transition. ETHICS AND DISSEMINATION: The trial has been granted ethical approval by the Medical Sciences Interdivisional Research Ethics Committee (IDREC) of the University of Oxford (Ref: R54329/RE001). All results originating from this study will be submitted for publication in scientific journals and presented at meetings and through the media. TRIAL REGISTRATION NUMBER: ISRCTN13180635;Pre-recruitment.

Deep divergence-based approach to clustering.

A promising direction in deep learning research consists in learning representations and simultaneously discovering cluster structure in unlabeled data by optimizing a discriminative loss function. As opposed to supervised deep learning, this line of research is in its infancy, and how to design and optimize suitable loss functions to train deep neural networks for clustering is still an open question. Our contribution to this emerging field is a new deep clustering network that leverages the discriminative power of information-theoretic divergence measures, which have been shown to be effective in traditional clustering. We propose a novel loss function that incorporates geometric regularization constraints, thus avoiding degenerate structures of the resulting clustering partition. Experiments on synthetic benchmarks and real datasets show that the proposed network achieves competitive performance with respect to other state-of-the-art methods, scales well to large datasets, and does not require pre-training steps.