q-Calculus and the Symmetrized Radioactive Decay Law

The radioactive decay law was first formulated by Ernest Rutherford and Frederick Soddy in 1902. As a well-known law, one of its primary applications is to determine the dates of ancient specimens. The process is known as radiocarbon dating and is subjected to the known properties of radioactive nuclei. In this paper, we implement quantum calculus to express the solution of the radioactive decay equation in symmetrized q-exponential form. Also, we explore a q-analog of the decay constant using Tsallis logarithmic function for various miscellaneous q-values. Furthermore, the factor-label method was applied to our analysis to show that the correct units remained intact under the application of quantum calculus. In conclusion, our work suggests that a variation of the q-parameter was akin to the production of a new isotope for all q in (0,1); the superadditive regime.

Adversarial-Learned Loss for Domain Adaptation

Recently, remarkable progress has been made in learning transferable representation across domains. Previous works in domain adaptation are majorly based on two techniques: domain-adversarial learning and self-training. However, domain-adversarial learning only aligns feature distributions between domains but does not consider whether the target features are discriminative. On the other hand, self-training utilizes the model predictions to enhance the discrimination of target features, but it is unable to explicitly align domain distributions. In order to combine the strengths of these two methods, we propose a novel method called Adversarial-Learned Loss for Domain Adaptation (ALDA). We first analyze the pseudo-label method, a typical self-training method. Nevertheless, there is a gap between pseudo-labels and the ground truth, which can cause incorrect training. Thus we introduce the confusion matrix, which is learned through an adversarial manner in ALDA, to reduce the gap and align the feature distributions. Finally, a new loss function is auto-constructed from the learned confusion matrix, which serves as the loss for unlabeled target samples. Our ALDA outperforms state-of-the-art approaches in four standard domain adaptation datasets. Our code is available at https://github.com/ZJULearning/ALDA.

Locating Multiple Sources of Contagion in Complex Networks under the SIR Model

Simultaneous outbreaks of contagion are a great threat against human life, resulting in great panic in society. It is urgent for us to find an efficient multiple sources localization method with the aim of studying its pathogenic mechanism and minimizing its harm. However, our ability to locate multiple sources is strictly limited by incomplete information about nodes and the inescapable randomness of the propagation process. In this paper, we present a valid approach, namely the Potential Concentration Label method, which helps locate multiple sources of contagion faster and more accurately in complex networks under the SIR(Susceptible-Infected-Recovered) model. Through label assignment in each node, our aim is to find the nodes with maximal value after several iterations. The experiments demonstrate that the accuracy of our multiple sources localization method is high enough. With the number of sources increasing, the accuracy of our method declines gradually. However, the accuracy remains at a slight fluctuation when average degree and network scale make a change. Moreover, our method still keeps a high multiple sources localization accuracy with noise of various intensities, which shows its strong anti-noise ability. I believe that our method provides a new perspective for accurate and fast multi-sources localization in complex networks.

Is More Better? Insights on Consumers’ Preferences for Nutritional Information on Wine Labelling

Background: Nowadays there is a strong debate on the need to introduce mandatory nutritional information on alcoholic beverages labels, and particularly on wine, as a tool to promote more health-conscious drinking patterns in society. In 2018, the European alcoholic beverages industry presented a self-regulatory proposal, now under assessment by the European Commission. The most critical issue is how to convey nutritional information to consumers, as producers should decide to apply information on label or off-label. Method: The current study measured, through a non-hypothetical, incentive compatible artefactual field experiment, Italian wine consumers (N = 103) preferences for four different formats of wine nutritional labelling, namely: back label with the indication of kcal for glass of wine, with the nutritional panel referred to 100 mL, without nutritional information (but with a link to an external website) and with the indication of key nutrients for glass of wine. Results: Findings reveal that respondents preferred the nutritional panel on the back label, assigning the lowest preference to the less informative wine label (only with a website recall). Furthermore, results show a low level of respondents’ knowledge of wine nutritional properties. Conclusion: Findings, while limited in terms of sample representativeness, seem to support the European Consumer Organisation and the European Alcohol Policy Alliance objection to an off-line label and the advocacy for a traditional and complete on label nutritional information on wine.