Search

The speech emotion recognition based on the deep networks on small samples is often a very challenging problem in natural language processing. The massive parameters of a deep network are much difficult to be trained reliably on small-quantity speech samples. Aiming at this problem, we propose a new method through the systematical cooperation of Generative Adversarial Network (GAN) and Long Short Term Memory (LSTM). In this method, it utilizes the adversarial training of GAN's generator and discriminator on speech spectrogram images to implement sufficient sample augmentation. A six-layer convolution neural network (CNN), followed in series by a two-layer LSTM, is designed to extract features from speech spectrograms. For accelerating the training of networks, the parameters of discriminator are transferred to our feature extractor. By the sample augmentation, a well-trained feature extraction network and an efficient classifier could be achieved. The tests and comparisons on two publicly available datasets, i.e., EMO-DB and IEMOCAP, show that our new method is effective, and it is often superior to some state-of-the-art methods. [ABSTRACT FROM AUTHOR]

Boiled silkworm cocoons have been used to treat 'Xiaoke disease' (diabetes mellitus) recorded in Chinese medicine for over 800 years. In recent years, it has been found that the active substance silk sericin (SS) has therapeutic benefits in treating type 2 diabetes mellitus (T2DM). SS promotes pancreatic islet signalling, the proliferation of pancreatic islet cells, and insulin secretion. It is inferred that SS enters the bloodstream after oral administration and plays a role in the body's circulation. As a natural protein, SS needs to resist digestion by proteases in the gastrointestinal tract and cross the gastrointestinal barrier after oral administration. It is currently unclear how SS crosses the gastrointestinal barrier and whether it exerts therapeutic effects on T2DM by entering the circulation. To study how SS crosses the gastrointestinal barrier and whether it enters the body circulation to exert a therapeutic effect on T2DM. SS was extracted from silkworm cocoons using an alkaline method with sodium carbonate. The antidigestive capacity of SS was detected using SDS-PAGE gel electrophoresis experiments. The mode of uptake and translocation of orally consumed SS in vivo was analysed using the AP-side to BL-side and BL-side-AP-side translocations, apparent Permeability coefficient (Papp), and Exocytosis rates (ER). The study compared the differences between the adSS group and the adSS + EDTA group by using Ethylenediaminetetraacetic acid (EDTA) to separate the tight junctions between Caco-2 cells. The aim was to analyze whether the transport mode of oral filaggrin proteins in vivo could be absorbed by bypass transport. By administering SS through oral and intraperitoneal injection to type 2 diabetic mice, we measured its concentration in the blood, as well as blood glucose and insulin levels, to determine its effectiveness in treating diabetes and its ability to enter the body's circulation for treatment. The molecular weight of SS decreased from 10k∼25 kDa to 10k∼15 kDa after in vitro simulated gastrointestinal fluid digestion, indicating its good antidigestive properties. The apparent Papp was greater than 1 × 10−6 cm·s-1, and the ER was between 0.5 and 1.5, indicating that adSS was well-absorbed and mainly passively transported. The Caco-2 cell model showed that the addition of EDTA promoted the transport of adSS, resulting in significantly larger Papp and ER values, indicating that adSS was absorbed by bypass transport. After oral administration of SS, the concentration of SS in the blood was lower than after intraperitoneal injection, which is 60% of intraperitoneal administration. Mice with a T2DM model who were administered SS for 5 weeks showed significant improvement in insulin resistance and glucose tolerance. Additionally, the pancreatic tissue appeared more regular. In the treatment of T2DM, injections of SS have been shown to be more effective than oral administration. Both oral and intraperitoneal injections have been partially involved in the circulation. SS is enzymatically cleaved by proteolytic enzymes in the gastrointestinal tract. The smaller molecules are partially absorbed into the body's circulation through passive and paracrine transport, exerting a therapeutic effect on T2DM. [Display omitted] [ABSTRACT FROM AUTHOR]

Neural networks, especially the latest deep learning, have exhibited good ability in estimating surface parameters from satellite remote sensing. However, thorough examinations of neural networks in the estimation of land surface temperature (LST) from satellite passive microwave (MW) observations are still lacking. Here, we examined the performances of the traditional neural network (NN), deep belief network (DBN), and convolutional neural network (CNN) in estimating LST from the AMSR-E and AMSR2 data over the Chinese landmass. The examinations were based on the same training set, validation set, and test set extracted from 2003, 2004, and 2009, respectively, for AMSR-E with a spatial resolution of 0.25°. For AMSR2, the three sets were extracted from 2013, 2014, and 2016 with a spatial resolution of 0.1°, respectively. MODIS LST played the role of "ground truth" in the training, validation, and testing. The examination results show that CNN is better than NN and DBN by 0.1–0.4 K. Different combinations of input parameters were examined to get the best combinations for the daytime and nighttime conditions. The best combinations are the brightness temperatures (BTs), NDVI, air temperature, and day of the year (DOY) for the daytime and BTs and air temperature for the nighttime. By adding three and one easily obtained parameters on the basis of BTs, the accuracies of LST estimates can be improved by 0.8 K and 0.3 K for the daytime and nighttime conditions, respectively. Compared with the MODIS LST, the CNN LST estimates yielded root-mean-square differences (RMSDs) of 2.19–3.58 K for the daytime and 1.43–2.14 K for the nighttime for diverse land cover types for AMSR-E. Validation against the in-situ LSTs showed that the CNN LSTs yielded root-mean-square errors of 2.10–4.72 K for forest and cropland sites. Further intercomparison indicated that ~50% of the CNN LSTs were closer to the MODIS LSTs than ESA's GlobTemperature AMSR-E LSTs, and the average RMSDs of the CNN LSTs were less than 3 K over dense vegetation compared to NASA's global land parameter data record air temperatures. This study helps better the understanding of the use of neural networks for estimating LST from satellite MW observations. [ABSTRACT FROM AUTHOR]