Generative Adversarial Networks (GANs) have gained a lot of attention from machine learning community due to their ability to learn and mimic an input data distribution. GANs consist of a discriminator and a generator working in tandem playing a min-max game to learn a target underlying data distribution; when fed with data-points sampled from a simpler distribution (like uniform or Gaussian distribution). Once trained, they allow synthetic generation of examples sampled from the target distribution. We investigate the application of GANs to generate synthetic feature vectors used for speech emotion recognition. Specifically, we investigate two set ups: (i) a vanilla GAN that learns the distribution of a lower dimensional representation of the actual higher dimensional feature vector and, (ii) a conditional GAN that learns the distribution of the higher dimensional feature vectors conditioned on the labels or the emotional class to which it belongs. As a potential practical application of these synthetically generated samples, we measure any improvement in a classifieru0027s performance when the synthetic data is used along with real data for training. We perform cross-validation analyses followed by a cross-corpus study.
The protein tyrosine phosphatases PTP1B and SHP2 are promising drug targets in treatment design for breast cancer. Searching for specific inhibitors of their activity has recently become the challenge of many studies. Previous work has indicated that the promising PTP inhibitors may be small compounds that are able to bind and interact with amino residues from the binding site.
Curcuminoids (Curs), oleoresins from Curcuma longa L., have known anticarcinogenic and anti-inflammatory properties, but high toxicity, poor aqueous solubility and susceptibility to degradation in body fluids are deterrents to their clinical administration. Poly(methyl methacrylate) nanoparticles (PMMA-NPs) are biocompatible and resilient and can entrap hydrophobic drugs. The present investigation is related to solubilizing Curs by incorporating them in these nanoparticles (NPs) and is related to a study comparing the anticarcinogenic effect of drug-loaded NPs with free Cur using lung cancer (A549) cell line. Freshly extracted oleoresins were post loaded in PMMA-NPs prepared using emulsion polymerization. The presence of the three components of oleoresins was confirmed by thin-layer chromatography. The size and morphology of void and loaded NPs were determined by dynamic light scattering, scanning electron microscopy and transmission electron microscopy. The NPs were spherical with diameters of 192.66±5 nm (void) and 199.16±5 nm (loaded). Drug loading and encapsulation efficiency were 6% and 93%, respectively. From the Fourier transform infrared spectroscopy spectra, the characteristic absorption vibration of poly(methyl methacrylate) and the bands at 1,383, 1,233 and 962 cm -1 for Cur moiety were observed. Drug release up to 10 days was estimated in buffer, saline and serum. The highest release of ~55% in ~3 days was noted in buffer that exhibited the highest bioavailability. The in vitro anticancer activity of loaded drug was evaluated up to 72 hours by MTT assay using A549 cell line. Cellular uptake of dye-loaded NPs was visualized within 30 minutes of incubation. The results revealed that the dose- and time-dependent cell death in case of loaded PMMA-NPs was comparable to that of free Cur. According to the study, the drug-loaded PMMA-NPs appear to be highly suitable for effective, localized and safe chemotherapy.