Ti2Nb10O29 is fabricated directly by solid-state reaction from commercial TiO2 and Nb2O5. Without further modification, the bulk anode exhibits a reversible capacity of 144 mA h g−1 at 10 C after 800 cycles. More impressively, Ti2Nb10O29/LiFePO4 full-cell 1 stabilizes 100 1000

Social media data with geotags can be used to track people's movements in their daily lives. By providing both rich text and movement information, visual analysis on social interesting challenging. In contrast traditional data, the sparseness irregularity of increase difficulty extracting patterns. To facilitate understanding movements, we present an interactive analytics system support exploration sparsely sampled trajectory from media. We propose a heuristic model reduce uncertainty caused...

For high-dimensional data, this work proposes two novel visual exploration methods to gain insights into the data aspect and dimension of data. The first is a Dimension Projection Matrix, as an extension scatterplot matrix. In matrix, each row or column represents group dimensions, cell shows projection (such MDS) with corresponding dimensions. second Tree, where every node either plot Matrix. Nodes are connected links child in tree covers subset parent node's dimensions items. While nodes...

An Sb/NPC composite with Sb nanoparticles uniformly dispersed in 1-D N-doped porous carbon exhibited enhanced cycle stability LIBs and NIBs.

Nanoporous silicon has been prepared through the air-oxidation demagnesiation of Mg₂Si at 600 °C for 10 hours (Mg₂Si + O₂→ Si MgO), followed by HCl washing.

Fe3O4 nanoflakes in an N-doped carbon matrix (Fe3O4 NF@NC) were prepared by solvothermal synthesis of and situ polymerization pyrrole on the surface followed heat treatment. The NF@NC is composed with a width 50–60 nm thickness 10 dispersed matrix. content varies from 18% to 50% controlling amount added, therefore 44% performs best. Due cooperation two-dimensional (2D) structure matrix, obtained (44% content) exhibits electrochemical performance reversible capacity 1046 mA h g−1 at 0.2 C (1...

Recently, researchers have started decomposing deep neural network models according to their semantics or functions. Recent work has shown the effectiveness of decomposed functional blocks for defending adversarial attacks, which add small input perturbation image fool DNN models. This proposes a profiling-based method decompose different blocks, lead effective path as new approach exploring DNNs' internal organization. Specifically, per-image can be aggregated class-level path, through we...

Abstract Ferric chloride‐graphite intercalation compounds (FeCl 3 –GICs) with stage 1 and 2 structures were synthesized by reacting FeCl expanded graphite (EG) in air a stainless‐steel autoclave. As rechargeable Li‐ion batteries, these –GICs exhibit high capacity, excellent cycling stability, superior rate capability, which could be attributed to their unique features. This work may enable new possibilities for the fabrication of batteries.

Speech separation based on deep neural networks (DNNs) has been widely studied recently, and achieved considerable success. However, previous studies are mostly fully-connected networks. In order to capture the local information of speech signals, we propose use convolutional maxout (CMNNs) separate noise by estimating ideal ratio mask time-frequency units. our work proposed CMNN is applied in frequency domain. By using filtering max-pooling, can model structure signals. Instead sigmoid...

Visualization and interactive analysis can help network administrators security analysts analyze the flow log data. The complexity of such an requires a combination knowledge experience from more domain experts to solve difficult problems faster with higher reliability. We developed online visual system called OCEANS address this topic by allowing close collaboration among create deeper insights in detecting events. Loading heterogeneous data source (netflow, IPS host status log), provides...

Aldo-keto reductases (AKRs) are a superfamily of enzymes that reduce aldehydes and ketones, have broad range substrates. An AKR gene, sakR1, was identified in the cyanobacterium Synechococcus sp. PCC 7002. A mutant strain with sakR1 inactivated sensitive to glycerol, carbon source can support heterotrophic growth It found null accumulated more toxic methylglyoxal than wild-type when glycerol added medium, suggesting SakR1 is involved detoxification methylglyoxal, highly metabolite damage...

Movie and TV subtitles are frequently employed in natural language processing (NLP) applications, but there limited Japanese-Chinese bilingual corpora accessible as a dataset to train neural machine translation (NMT) models. In our previous study, we effectively constructed corpus of considerable size containing text data both Japanese Chinese by collecting subtitle from websites that host movies television series. The unsatisfactory performance the initial corpus, Web-Crawled Corpus (WCC-JC...

With the proliferation of digital cameras and mobile devices, people are taking many more photos than ever before. The explosive growth personal leads to problems photo organization management. There is a growing need for tools automatically manage collections. Recognizing events in collections one efficient way organize photos. use textual event labels can allow us categorize locate an without browsing through entire collection. Most existing research on this topic focuses recognizing from...

People are taking more photos than ever before in recent years. To effectively organize these personal photos, the usually assigned to albums according their events. An efficient way manage our would be if we could recognize events of automatically. In this paper, study problem recognizing photo albums. Recognizing is a new challenge since contents complicated traditional single-photo tasks, not all an album relevant event and single often fails convey meaningful semantic behind album. solve...

Social networks nowadays have become an important form of communication in which users can post their current status or share lives by mobile phones the Web. In this paper, we develop effective and efficient model to estimate continuous tie strength between for friend recommendation with heterogeneous data from social media community. We categorize those multimodal into two classes: interaction (e.g., comments, marking favorite photos) similarity data(e.g., common friends, groups, tags, geo,...