Novel zwitterionic polyurethanes containing sulfobetaines, featuring excellent multi-shape-memory properties and self-healing properties, are prepared from N-methyldiethanolamine, hexamethylene diisocyanate and 1,3-propanesultone.

Like other problems in computer vision, offline handwritten Chinese character recognition (HCCR) has achieved impressive results using convolutional neural network (CNN)-based methods. However, larger and deeper networks are needed to deliver state-of-the-art results in this domain. Such networks intuitively appear to incur high computational cost, and require the storage of a large number of parameters, which renders them unfeasible for deployment in portable devices. To solve this problem, we propose a Global Supervised Low-rank Expansion (GSLRE) method and an Adaptive Drop-weight (ADW) technique to solve the problems of speed and storage capacity. We design a nine-layer CNN for HCCR consisting of 3,755 classes, and devise an algorithm that can reduce the networks computational cost by nine times and compress the network to 1/18 of the original size of the baseline model, with only a 0.21% drop in accuracy. In tests, the proposed algorithm surpassed the best single-network performance reported thus far in the literature while requiring only 2.3 MB for storage. Furthermore, when integrated with our effective forward implementation, the recognition of an offline character image took only 9.7 ms on a CPU. Compared with the state-of-the-art CNN model for HCCR, our approach is approximately 30 times faster, yet 10 times more cost efficient.

Although phase transformation is suggested as a key step in biomineralization, the chemical scenario about how organic molecules mediate inorganic phase transformations is still unclear. The inhibitory effect of amino acids on hydroxyapatite (HAP, the main inorganic component of biological hard tissues such as bone and enamel) formation was concluded by the previous biomimetic modeling based upon direct solution crystallization. Here we demonstrate that acidic amino acids, Asp and Glu, could promote HAP crystallization from its precursor crystal, brushite (DCPD). However, such a promotion effect could not be observed when the nonacidic amino acids were applied in the transformation-based HAP formation. We found that the specific modification of acidic amino acid on crystal-solution interfaces played a key role in the phase transition. The distinct properties between DCPD and HAP in the solution resulted in an interfacial energy barrier to suppress the spontaneous formation of HAP phase on DCPD phase. Different from the other amino acids, the carboxylate-rich amino acids, Asp and Glu, could modify the interfacial characteristics of these two calcium phosphate crystals to make them similar to each other. The experiments confirmed that the involvement of Asp or Glu reduced the interfacial energy barrier between DCPD and HAP, leading to a trigger effect on the phase transformation. An in-depth understanding about the unique roles of acidic amino acids may contribute to understanding phase transformation controls druing biomineralization.