The cosmetic use of bleaching products in Dakar, Senegal: socio-economic factors and claimed motivations In order to evaluate the reasons for the cosmetic use of skin bleaching products, a questionnaire was proposed to a sample of 368 women presenting at a dermatological centre in Dakar. According to the collected data, bleaching products appear to be more frequently used among women from Dakar, aged from 20 to 40 years, married or having a boy- friend, with an educational level equivalent to primary school, a job implying contacts with customers, and among women having access to certain consumption goods (TV, phone, car). According to their statements, the main motivations of women using bleaching products are the following: be fashionable, be pretty, imitate close friends, practice self- medication. Female friends of bleaching products users are said to often compliment and prompt them, contrarily to family members who generally criticize strongly the practice; the husbands present an intermediary position. These data suggest that bleaching products' s use works like a fashion phenomenon, with both identification to a group (mainly, female friends) and a claim to certain values, especially urbanity, modernity, adult femininity, power of seduction, and access to a certain social level. A wish of emancipation toward classical patterns of women seems present. There is a significant ambivalence of bleaching products users concerning their cosmetic practices. According to thèse data, the usual explanations for the cosmetic use of bleaching products, i. e. a negative perception of black skin making reference to racial typology, are not confirmed. In order to understand the current use of bleaching products, it is essential to adopt a historical perspective, especially with regard to local perceptions of skin colour differences, as to local strategies of seduction.

This paper revisit the methodology of system identification and shows how new paradigms from machine learning can be used to improve the model identification performance in the case of non-linear systems observed with noisy and unbalanced dataset. We prove that using importance sampling schemes in system identification can provide significant performance boost on a wide variety of systems, in particular when some of the system dynamic is only exhibited by relatively rare events. The performance of the approaches is evaluated on a real and simulated drone and two standard datasets from real robotic systems. Our approach consistently outperforms baseline approaches on these datasets, all the more when the datasets are noisy and unbalanced.

Machine learning allows to create complex models if provided with enough data, hence challenging more traditional system identification methods. We compare the quality of neural networks and an ARX model when use in an model predictive control to command a drone in a simulated environment. The training of neural networks can be challenging when the data is scarce or datasets are unbalanced. We propose an adaptation of prioritized replay to system identification in order to mitigate these problems. We illustrate the advantages and limits of this training method on the control task of a simulated drone.Index Terms-identification, model predictive control, neural networks, learning