It has been shown that news events influence the trends of stock price movements. However, previous work on news-driven stock market prediction rely on shallow features (such as bags-of-words, named entities and noun phrases), which do not capture structured entity-relation information, and hence cannot represent complete and exact events. Recent advances in Open Information Extraction (Open IE) techniques enable the extraction of structured events from web-scale data. We propose to adapt Open IE technology for event-based stock price movement prediction, extracting structured events from large-scale public news without manual efforts. Both linear and nonlinear models are employed to empirically investigate the hidden and complex relationships between events and the stock market. Largescale experiments show that the accuracy of S&P 500 index prediction is 60%, and that of individual stock prediction can be over 70%. Our event-based system outperforms bags-of-words-based baselines, and previously reported systems trained on S&P 500 stock historical data.

DNA self-assembling nanostructure has been considered as a promising candidate as a drug delivery vehicle because of its compactness, mechanical stability, and noncytotoxicity. In this work, we developed functional, multiform DNA nanostructures by appending a tumor-penetrating peptide to tetrahedral DNA nanostructure (p-TDN). This functional structure is able to efficiently increase the rate of uptake of glioblastoma cell U87MG compared with the DNA tetrahedron and the double-stranded DNA structures. We found that the DNA tetrahedron plays the main role in the endocytosis of U87MG cells, whereas the tumor-penetrating peptide could also bind to transmembrane glycoprotein neuropilin-1 and mediate the endocytosis of the p-TDN nanostructure. Moreover, given the high efficiency of the growth inhibitory effect of the p-TDN loading doxorubicin hydrochloride, the p-TDN distinguishes itself as a promising candidate as an effective delivery carrier.

To meet the rapidly growing demand, it is necessary to develop novel flexible energy storage devices with a high energy density in a limited area, a fast charging ability, a low cost for mass production and a miniaturized device size. To address the above issues, here we introduce the co-electro-deposition strategy, which is able to prepare an electrode material with a high areal capacitance (1670 mF cm À2 at 0.5 mA cm À2 ), a high areal mass (8.5 mg cm À2 ), an excellent mechanical robustness, a high through-put and great convenience even on a piece of a ubiquitous stainless steel mesh current collector. Based on this advancement, we are able to obtain an ultrathin (less than 200 mm) aqueous asymmetric supercapacitor device with a high energy density (1.8 Â 10 À3 W h cm À3 ), a high power density (0.38 W cm À3 at 3.62 Â 10 À4 W h cm À3 ) and an excellent rate capability. This energy storage device is integrated into a prototype smart card to drive a light emitting diode (LED) indicator, which is charged for 5 seconds and can light up the indicator for more than 2 hours, demonstrating great promise in miniaturized novel flexible energy storage devices.

No abstract

No abstract

No abstract

No abstract

In this paper, we focus on the personalized response generation for conversational systems. Based on the sequence to sequence learning, especially the encoder-decoder framework, we propose a two-phase approach, namely initialization then adaptation, to model the responding style of human and then generate personalized responses. For evaluation, we propose a novel human aided method to evaluate the performance of the personalized response generation models by online real-time conversation and offline human judgement. Moreover, the lexical divergence of the responses generated by the 5 personalized models indicates that the proposed two-phase approach achieves good results on modeling the responding style of human and generating personalized responses for the conversational systems.