To save content items to your account,
please confirm that you agree to abide by our usage policies.
If this is the first time you use this feature, you will be asked to authorise Cambridge Core to connect with your account.
Find out more about saving content to .
To save content items to your Kindle, first ensure firstname.lastname@example.org
is added to your Approved Personal Document E-mail List under your Personal Document Settings
on the Manage Your Content and Devices page of your Amazon account. Then enter the ‘name’ part
of your Kindle email address below.
Find out more about saving to your Kindle.
Note you can select to save to either the @free.kindle.com or @kindle.com variations.
‘@free.kindle.com’ emails are free but can only be saved to your device when it is connected to wi-fi.
‘@kindle.com’ emails can be delivered even when you are not connected to wi-fi, but note that service fees apply.
Conversational recommender system (CRS) needs to be seamlessly integrated between the two modules of recommendation and dialog, aiming to recommend high-quality items to users through multiple rounds of interactive dialogs. Items can typically refer to goods, movies, news, etc. Through this form of interactive dialog, users can express their preferences in real time, and the system can fully understand the user’s thoughts and recommend corresponding items. Although mainstream dialog recommendation systems have improved the performance to some extent, there are still some key issues, such as insufficient consideration of the entity’s order in the dialog, the different contributions of items in the dialog history, and the low diversity of generated responses. To address these shortcomings, we propose an improved dialog context model based on time-series features. Firstly, we augment the semantic representation of words and items using two external knowledge graphs and align the semantic space using mutual information maximization techniques. Secondly, we add a retrieval model to the dialog recommendation system to provide auxiliary information for generating replies. We then utilize a deep timing network to serialize the dialog content and more accurately learn the feature relationship between users and items for recommendation. In this paper, the dialog recommendation system is divided into two components, and different evaluation indicators are used to evaluate the performance of the dialog component and the recommendation component. Experimental results on widely used benchmarks show that the proposed method is effective.
A 6-week growth trial was conducted to evaluate the influences of dietary valine (Val) levels on growth, protein utilisation, immunity, antioxidant status and gut micromorphology of juvenile hybrid groupers. Seven isoenergetic, isoproteic and isolipidic diets were formulated to contain graded Val levels (1·21, 1·32, 1·45, 1·58, 1·69, 1·82 and 1·94 %, DM basis). Each experimental diet was hand-fed to triplicate groups of twelve hybrid grouper juveniles. Results showed that weight gain percentage (WG%), protein productive value, protein efficiency ratio, and feed efficiency were increased as dietary Val level increased, reaching a peak value at 1·58 % dietary Val. The quadratic regression analysis of WG% against dietary Val levels indicated that the optimum dietary Val requirement for hybrid groupers was estimated to be 1·56 %. Gut micromorphology and expression of growth hormone in pituitary, insulin-like growth factor 1, target of rapamycin and S6 kinase 1 in liver were significantly affected by dietary Val levels. In serum, fish fed 1·58 % dietary Val had higher superoxide dismutase, catalase, lysozyme activities and IgM concentrations than fish fed other dietary Val levels. Fish fed 1·58 % dietary Val had higher expression of NF-E2-related factor 2 in head kidney than fish fed other dietary Val levels. Generally, the optimum dietary Val requirement for maximal growth of hybrid groupers was estimated to be 1·56 % of DM, corresponding to 3·16 % of dietary protein, and dietary Val levels affected growth, protein utilisation, immunity and antioxidant status in hybrid groupers.
Deep learning methods have been applied to randomly generate images, such as in fashion, furniture design. To date, consideration of human aspects which play a vital role in a design process has not been given significant attention in deep learning approaches. In this paper, results are reported from a human- in-the-loop design method where brain EEG signals are used to capture preferable design features. In the framework developed, an encoder extracting EEG features from raw signals recorded from subjects when viewing images from ImageNet are learned. Secondly, a GAN model is trained conditioned on the encoded EEG features to generate design images. Thirdly, the trained model is used to generate design images from a person's EEG measured brain activity in the cognitive process of thinking about a design. To verify the proposed method, a case study is presented following the proposed approach. The results indicate that the method can generate preferred designs styles guided by the preference related brain signals. In addition, this method could also help improve communication between designers and clients where clients might not be able to express design requests clearly.
Phase separation of InxGa1−xN into Ga-rich and In-rich regions has been studied by electron energy-loss spectroscopy (EELS) in a monochromated, aberration corrected scanning transmission electron microscope (STEM). We analyze the full spectral information contained in EELS of InGaN, combining for the first time studies of high-energy and low-energy ionization edges, plasmon, and valence losses. Elemental maps of the N K, In M4,5 and Ga L2,3 edges recorded by spectrum imaging at 100 kV reveal sub-nm fluctuations of the local indium content. The low energetic edges of Ga M4,5 and In N4,5 partially overlap with the plasmon peaks. Both have been fitted iteratively to a linear superimposition of reference spectra for GaN, InN, and InGaN, providing a direct measurement of phase separation at the nm-scale. Bandgap measurements are limited in real space by scattering delocalization rather than the electron beam size to ∼10 nm for small bandgaps, and their energetic accuracy by the method of fitting the onset of the joint density of states rather than energy resolution. For an In0.62Ga0.38N thin film we show that phase separation occurs on several length scales.
Phase separation of InxGa1-xN alloys into Ga-rich and In-rich regions was observed by a number of research groups for samples grown with high indium content, x. Due to the radiation sensitivity of InGaN to beam damage by fast electrons, high-resolution imaging in transmission electron microscopy (TEM) or core-loss electron energy-loss spectroscopy (EELS) may lead to erroneous results. Low-loss EELS can yield spectra of the plasmon loss regions at much lower electron fluxes. Unfortunately, due to their delayed edge onset, the low energetic core losses of Ga and In partially overlap with the plasmon peaks, all of which shift with indium content.
Here we demonstrate a method to quantify phase separation in InGaN thin films from the low-loss region in EELS by simultaneously fitting both plasmon and core losses over the energy range of 13-30eV. Phase separation is shown to lead to a broadening of the plasmon peak and the overlapping core losses, resulting in an unreliable determination of the indium concentration from analyzing the plasmon peak position alone if phase separation is present. For x=0.3 and x=0.59, the relative contributions of the binary compounds are negligibly small and indicate random alloys. For xnom.=0.62 we observed strong broadening, indicating phase separation.
Rapid growth processing of KDP crystals was improved by employing continuous filtration to eliminate bulk defects. The performances of the KDP crystals, including scattering defects, laser damage resistance and transmittance, were measured and analyzed. Compared with rapid-grown KDP without continuous filtration, the transmittance in the near-infrared was increased by at least 2%, almost all of ‘micron size’ defects were eliminated and ‘sub-micron size’ defects were decreased by approximately 90%. Laser damage testing revealed that the laser-induced damage thresholds (LIDTs), as well as the consistency of the LIDTs from sample to sample, were improved greatly. Moreover, it identified that ‘micron size’ defects were the precursors which initiated laser damage at relative lower laser fluence (4–6 J cm−2), and there was a lower correlation between smaller size scattering defects and laser damage initiation. The improved consistency in the LIDTs, attributed to elimination of ‘micron size’ defects, and LIDT enhancement originated from the decreased absorption of the KDP crystals.
In this paper we make an empirical comparison of sales time series for online and offline channels. In particular, we analyse the sales dynamic and fluctuation level underlying the sales time series in different channels. The accumulative daily sales distributions of commodities are analysed statistically and the daily sales series are also studied from the perspective of complex networks. We find that most of the commodities' accumulative sales distributions can be fitted by power-law distributions. Visibility graphs are constructed for the daily sales series, and the accumulative degree distributions are also investigated – it is found that they also almost follow power-law distribution. The constant parameter α indicates that different specifications of the same goods have different sales characteristics, and different forms of packaging of commodities, either special offer or ordinary, also show distinctive sales fluctuation levels. The differences show that the direction of these relationships is opposite for online and offline channels.
Email your librarian or administrator to recommend adding this to your organisation's collection.