We propose a new framework to rank image attractiveness using a novel pairwise deep network trained with a large set of side-by-side multi-labeled image pairs from a web image index. The judges only provide relative ranking between two images without the need to directly assign an absolute score, or rate any predefined image attribute, thus making the rating more intuitive and accurate. We investigate a deep attractiveness rank net (DARN), a combination of deep convolutional neural network and rank net, to directly learn an attractiveness score mean and variance for each image and the underlying criteria the judges use to label each pair. The extension of this model (DARN-V2) is able to adapt to individual judge's personal preference. We also show the attractiveness of search results are significantly improved by using this attractiveness information in a real commercial search engine. We evaluate our model against other state-ofthe-art models on our side-by-side web test data and another public aesthetic data set. With much less judgments (1M vs 50M), our model outperforms on side-by-side labeled data, and is comparable on data labeled by absolute score.

An analysis of the available automated indoor temperature control systems shows that all of them have a significant shortage. The systems do not provide for predictive temperature control in an automated form. Such changes are determined and made, depending on the qualification, by the operator based on personal experience and the reports received by him on the weather channels. Possible errors lead to loss of productivity or disruption of technological modes of production. (Research purpose) The research purpose is in developing a system of automated predictive control of the temperature regime of industrial premises to improve energy efficiency and the quality of maintaining the necessary climatic conditions. (Materials and methods) Increasing the efficiency of greenhouses can be achieved through the introduction of an automated and more economical temperature control system. The proposed device additionally includes a weather forecast receiving unit, made on the basis of microprocessor technology, for reliable signal reception. The air that circulates inside the greenhouse through heat exchange circuits was used as a heat carrier. (Results and discussion) An additional heat (cold) reserve is created in the greenhouse, which allows to maintain the set temperature in the room, taking into account the weather forecast. The results presented in patents No. 2710010, 2667684 and the implemented control system based on certificates of state registration of programs for electronic computers No. 2015618030, 20146115490, and 2014617605 exceed the results obtained by the authors of the proposed developments based on patents No. 2586923 and No. 80308. (Conclusions) The use of new technical solutions and technologies for providing the basis for the development of a new principle for the construction of automated control systems for the temperature regime of industrial premises, taking into account the predicted changes in external conditions. The practical result of the conducted research is: a concept, a control algorithm, a manufactured and tested experimental sample of a device for controlling the temperature regime in a greenhouse.