Matching images and sentences demands a fine understanding of both modalities. In this article, we propose a new system to discriminatively embed the image and text to a shared visual-textual space. In this field, most existing works apply the ranking loss to pull the positive image/text pairs close and push the negative pairs apart from each other. However, directly deploying the ranking loss on heterogeneous features (i.e., text and image features) is less effective, because it is hard to find appropriate triplets at the beginning. So the naive way of using the ranking loss may compromise the network from learning inter-modal relationship. To address this problem, we propose the instance loss, which explicitly considers the intra-modal data distribution. It is based on an unsupervised assumption that each image/text group can be viewed as a class. So the network can learn the fine granularity from every image/text group. The experiment shows that the instance loss offers better weight initialization for the ranking loss, so that more discriminative embeddings can be learned. Besides, existing works usually apply the off-the-shelf features, i.e., word2vec and fixed visual feature. So in a minor contribution, this article constructs an end-to-end dual-path convolutional network to learn the image and text representations. End-to-end learning allows the system to directly learn from the data and fully utilize the supervision. On two generic retrieval datasets (Flickr30k and MSCOCO), experiments demonstrate that our method yields competitive accuracy compared to state-of-the-art methods. Moreover, in language-based person retrieval, we improve the state of the art by a large margin. The code has been made publicly available.
Direct flame fuel cells (DFFCs) have been investigated as an alternative means of combustion based power generation devices, but current challenges for this technology have included low fuel utilization and efficiency. In order to overcome these obstacles a new micro-tubular flame-assisted fuel cell (mT-FFC) concept is developed in this work and its performance is assessed at different equivalence ratios and temperatures. The concept is based on fuel-rich combustion exhaust, with the combustion equivalence ratio controlled and the exhaust flowing through the fuel cell for complete electrochemical energy conversion. The results were compared to a hydrogen baseline with the same electron potential as the fuel-rich exhaust. The mT-FFC concept offers significant advantages including high fuel utilization and greater performance stability compared to DFFCs.
The performance of yttria-stabilized zirconia (YSZ)–samaria-doped ceria (SDC) dual layer electrolyte anode-supported solid oxide fuel cell (AS-SOFC) was investigated. Tape-casting, lamination, and co-sintering of the NiO–YSZ anode followed by wet powder spraying of the SDC buffer layer and BSCF cathode was proposed for fabrication of these cells as an effective means of reducing the number of sintering stages required. The AS-SOFC showed a significant fuel cell performance of ∼1.9 W cm−2 at 800 °C. The fuel cell performance varies significantly with the sintering temperature of the SDC buffer layer. An optimal buffer layer sintering temperature of 1350 °C occurs due to a balance between the YSZ–SDC contact and densification at low sintering temperature and reactions between YSZ and SDC at high sintering temperatures. At high sintering temperatures, the reactions between YSZ and SDC have a detrimental effect on the fuel cell performance resulting in no power at a sintering temperature of 1500 °C.
Interest in measurement of children's executive functions has shown a major increase over the past two decades. The present study evaluates two new apps (EYT and eFun) for measuring executive functions in children. The results of this study show that children (aged 5-8) enjoy executive function assessment in the form of games on an iPad. However, only one executive function, EYT working memory, showed significant positive correlations with several types of grades (e.g., English and maths) in primary school students. New, self-assessed, child-friendly executive function measurement tools have the potential to provide future possibilities for teachers to integrate information on cognitive ability into student learning plans.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.