No abstract

Abstract-As a promising paradigm to reduce both capital and operating expenditures, the cloud radio access network (C-RAN) has been shown to provide high spectral efficiency and energy efficiency. Motivated by its significant theoretical performance gains and potential advantages, C-RANs have been advocated by both the industry and research community. This paper comprehensively surveys the recent advances of C-RANs, including system architectures, key techniques, and open issues. The system architectures with different functional splits and the corresponding characteristics are comprehensively summarized and discussed. The state-of-the-art key techniques in C-RANs are classified as: the fronthaul compression, large-scale collaborative processing, and channel estimation in the physical layer; and the radio resource allocation and optimization in the upper layer. Additionally, given the extensiveness of the research area, open issues and challenges are presented to spur future investigations, in which the involvement of edge cache, big data mining, socialaware device-to-device, cognitive radio, software defined network, and physical layer security for C-RANs are discussed, and the progress of testbed development and trial test are introduced as well.

Linguistic knowledge is of great benefit to scene text recognition. However, how to effectively model linguistic rules in end-to-end deep networks remains a research challenge. In this paper, we argue that the limited capacity of language models comes from: 1) implicitly language modeling; 2) unidirectional feature representation; and 3) language model with noise input. Correspondingly, we propose an autonomous, bidirectional and iterative ABINet for scene text recognition. Firstly, the autonomous suggests to block gradient flow between vision and language models to enforce explicitly language modeling. Secondly, a novel bidirectional cloze network (BCN) as the language model is proposed based on bidirectional feature representation. Thirdly, we propose an execution manner of iterative correction for language model which can effectively alleviate the impact of noise input. Additionally, based on the ensemble of iterative predictions, we propose a self-training method which can learn from unlabeled images effectively. Extensive experiments indicate that ABINet has superiority on lowquality images and achieves state-of-the-art results on several mainstream benchmarks. Besides, the ABINet trained with ensemble self-training shows promising improvement in realizing human-level recognition. Code is available at https://github.com/FangShancheng/ABINet.

With the great success of pre-trained language models, the pretrain-finetune paradigm now becomes the undoubtedly dominant solution for natural language understanding (NLU) tasks. At the fine-tune stage, target task data is usually introduced in a completely random order and treated equally. However, examples in NLU tasks can vary greatly in difficulty, and similar to human learning procedure, language models can benefit from an easy-to-difficult curriculum. Based on this idea, we propose our Curriculum Learning approach. By reviewing the trainset in a crossed way, we are able to distinguish easy examples from difficult ones, and arrange a curriculum for language models. Without any manual model architecture design or use of external data, our Curriculum Learning approach obtains significant and universal performance improvements on a wide range of NLU tasks.

No abstract

Poly(acrylic acid) (PAA) nanogels with a hollow core-porous shell structure were prepared by the direct polymerization of an acrylic acid monomer in the presence of hydroxypropylcellulose (HPC) and a cross-linking agent, N,N-methylenebisacrylamide, followed by removal of HPC from the generated HPC-PAA nanoparticles in a basic environment. The properties of PAA nanogel were characterized by dynamic light scattering, FT-IR, transmission electron microscopy, and atomic force microscopy. It is found that the nanogels have a hollow core-porous shell structure. Protein, bovine serum albumin (BSA), and an antitumor agent, doxorubicin hydrochloride, were used as model drugs to investigate their loading abilities as versatile drug-delivery vehicles. The nanogel exhibits surprisingly high loading ability to both protein and small molecular drugs. For example, the maximum BSA loading capacity of PAA nanogel can reach as high as 800% (i.e., 1 mg of nanogel can load about 8.0 mg of BSA). This high loading capacity may be related with the hollow core-porous shell structure of PAA nanogels. PAA nanogels have also shown sustained drug release properties and can cross biological barriers to deliver loaded cargo inside cells. Considering the high stability of the materials, simple and mild preparation procedure, high loading capacity, sustained-release property, and ability to protect biological agents from denaturation, PAA nanogels should be promising drug-delivery carriers for drug-delivery systems.

To develop a novel type of semi-IPN particles using biocompatible materials, hydroxypropylcellulose-poly(acrylic acid) (HPC-PAA) particles with semi-interpenetrating polymer network structure and a porosity-structural surface were prepared by direct polymerization of acrylic acid monomer in the reaction system comprised of HPC and AA monomer and N,N'-methylenebisacrylamide (MBAAm). The properties of HPC-PAA gel particles were characterized by dynamic light scattering, FT-IR, transmission electron microscopy, and atomic force microscope. It is found that the formation of HPC-PAA gel particles is driven by the hydrogen bonding interaction between proton-donating PAA and proton-accepting HPC. These HPC-PAA gel particles exhibit thermo and pH dual-responsive behaviors. Depending on the chemical composition and the degree of cross-linking, the thermo-responsive property of HPC-PAA gel particles can be shifted from the UCST to the LCST property, and particle sizes can be changed from 100 to 1 microm in a controllable way. Successful loading of the gel particles with oxaliplatin, a hydrophilic antitumor drug, was achieved by take advantage of the complex interaction between the platinum atom of oxaliplatin and the carboxylic group of PAA in the gel particles. In vitro cytotoxicity assay indicates that the oxalipatin-loaded HPC-PAA gel particles have high anticancer activity. Considering the good biosafety, simple and mild preparation strategy and tunable size as well as the stimuli-responsive properties, the HPC-PAA gel particles should be a promising candidate for the drug delivery system.

No abstract