Search

Escaping of the liquid molecules from their liquid bulk into the vapour phase at the vapour-liquid interface is controlled by the vapour diffusion process, which nevertheless hardly senses the macroscopic shape of this interface. Here, deformed sessile drops due to gravity and surface tension with various interfacial profiles are realised by tilting flat substrates. The symmetry broken of the sessile drop geometry leads to a different evaporation behavior compared to a drop with a symmetric cap on a horizontal substrate. Rather than the vapour-diffusion mechanism, heat-diffusion regime is defined here to calculate the local evaporation flux along the deformed drop interface. A local heat resistance, characterised by the liquid layer thickness perpendicular to the substrate, is proposed to relate the local evaporation flux. We find that the drops with and without deformation evaporate with a minimum flux at the drop apex, while up to a maximum one with a significantly larger but finite value at the contact line. Counterintuitively, the deviation from the symmetric shape due to the deformation on a slope, surprisingly enhances the total evaporation rate; and the smaller contact angle, the more significant enhancement. Larger tilt quickens the overall evaporation process and induces a more heterogeneous distribution of evaporative flux under gravity. Interestingly, with this concept of heat flux, an intrinsic heat resistance is conceivable around the contact line, which naturally removes the singularity of the evaporation flux showing in the vapour-diffusion model. The detailed non-uniform evaporation flux suggests ways to control the self-assembly, microstructures of deposit with engineering applications particularly in three dimensional printing where drying on slopes is inevitable., Comment: 12 pages, 7 figures

In recent years, there are numerous works been proposed to leverage the techniques of deep learning to improve social-aware recommendation performance. In most cases, it requires a larger number of data to train a robust deep learning model, which contains a lot of parameters to fit training data. However, both data of user ratings and social networks are facing critical sparse problem, which makes it not easy to train a robust deep neural network model. Towards this problem, we propose a novel Correlative Denoising Autoencoder (CoDAE) method by taking correlations between users with multiple roles into account to learn robust representations from sparse inputs of ratings and social networks for recommendation. We develop the CoDAE model by utilizing three separated autoencoders to learn user features with roles of rater, truster and trustee, respectively. Especially, on account of that each input unit of user vectors with roles of truster and trustee is corresponding to a particular user, we propose to utilize shared parameters to learn common information of the units that corresponding to same users. Moreover, we propose a related regularization term to learn correlations between user features that learnt by the three subnetworks of CoDAE model. We further conduct a series of experiments to evaluate the proposed method on two public datasets for Top-N recommendation task. The experimental results demonstrate that the proposed model outperforms state-of-the-art algorithms on rank-sensitive metrics of MAP and NDCG., Comment: Accepted by Frontiers of Computer Science