Quercetin is one of the most abundant flavonoids and the defense secondary metabolites in plants. In this study, the effect of quercetin on the growth of the silkworm larvae was investigated. Cytochrome P450 monooxygenases (P450s), glutathione S-transferases (GSTs), and carboxylesterases (COE) were assayed after exposure to different concentrations of quercetin for 3 d (short-term) and 7 d (long-term), respectively. The results showed that the weight gain of the silkworm larvae significantly decreased after the larvae were treated by different concentrations of quercetin except for the treatment with 0.5% quercetin. Activities of P450, GST, and COE were induced by 0.5 or 1% concentration of quercetin. In the midgut, the induction activity of P450s was reached to the highest level (2.3-fold) by 1% quercetin for 7 d, the highest induction activities of GSTs toward CHP and CDNB were 4.1-fold and 2.6-fold of controls by 1% quercetin after 7 d exposure, respectively. For COEs, the highest activity (2.3-fold) was induced by 0.5% quercetin for 7 d. However, P450s in whole body were higher inducible activities in short-term treatment than those in long-term treatment. The responses of eight cytochrome P450 (CYP) genes belonged to CYP6 and CYP9 families and seven GST genes were detected with real-time polymerase chain reaction. In addition, the genes induced by quercetin significantly were confirmed by qRT-PCR. CYP6AB5, CYP6B29, and GSTe8 were identified as inducible genes, of which the highest induction levels were 10.9-fold (0.5% quercetin for 7 d), 6.2-fold (1% quercetin for 7 d), and 7.1-fold (1% quercetin for 7 d), respectively.

Our work focuses on tackling the challenging but natural visual recognition task of long-tailed data distribution (i.e., a few classes occupy most of the data, while most classes have rarely few samples). In the literature, class re-balancing strategies (e.g., re-weighting and re-sampling) are the prominent and effective methods proposed to alleviate the extreme imbalance for dealing with long-tailed problems. In this paper, we firstly discover that these rebalancing methods achieving satisfactory recognition accuracy owes to that they could significantly promote the classifier learning of deep networks. However, at the same time, they will unexpectedly damage the representative ability of the learned deep features to some extent. Therefore, we propose a unified Bilateral-Branch Network (BBN) to take care of both representation learning and classifier learning simultaneously, where each branch does perform its own duty separately. In particular, our BBN model is further equipped with a novel cumulative learning strategy, which is designed to first learn the universal patterns and then pay attention to the tail data gradually. Extensive experiments on four benchmark datasets, including the large-scale iNaturalist ones, justify that the proposed BBN can significantly outperform state-of-the-art methods. Furthermore, validation experiments can demonstrate both our preliminary discovery and effectiveness of tailored designs in BBN for long-tailed problems. Our method won the first place in the iNaturalist 2019 large scale species classification competition, and our code is open-source and available at https://github.com/Megvii-Nanjing/BBN . * Q. Cui and Z.-M. Chen's contribution was made when they were interns in Megvii Research Nanjing, Megvii Technology, China. X.

In this paper, a band-pass filter based on half-mode substrate integrated waveguide (HMSIW) and double-layer spoof surface plasmon polaritons (SSPPs) consisting of two corrugated metal strips is proposed, which can realize band-pass transmission by etching periodic grooves at the top and bottom metal layers of the HMSIW. Moreover, the influences of important parameters on the performance of the proposed band-pass filter are analyzed by parametric study. By changing the key parameters, the low and high cut-off frequency can be controlled independently. The corresponding equivalent circuit of the proposed band-pass filter is put forward to explain the physical mechanism. Compared with the previous structures, this structure features smaller size, wider bandwidth and lower loss. Simulated results show that the proposed band-pass filter achieves a bandwidth (for |S11| < −10 dB and |S21| > −0.8 dB) of about 69.77% (15.6–32.1 GHz). The measured results have good agreements with the simulated ones, which verify that the proposed band-pass filter has good performances and potential applications at the microwave frequencies.

A compact wideband dual-polarized omnidirectional antenna with good isolation for the fifth generation/wireless local area networks communication is presented in this paper. The dual-polarized characteristic of the proposed antenna is achieved by combining a vertical polarization (VP) monopole antenna and horizontal polarization (HP) cross bow-tie antenna, which is excited by a 50 Sub-Miniature-A (SMA) connector and a broadband feeding network, respectively. The size of the proposed antenna is only 0.38λ×0.38λ×0.27λ (with λ being the wavelength of the lowest frequency). The simulated isolation of the proposed dual-polarized omnidirectional antenna is below −38dB in the bandwidth from 2.3 GHz to 3.7 GHz (about 46.7%, |S 11 | < −10 dB), taking no account of the commercial RF-Balun and SMA connectors. Moreover, the measured isolation of the proposed antenna with a feeding network is about −20 dB. The gain variations at the center frequency in the horizontal plane are 1.6 dB for VP element and 2.4 dB for HP element. And the lowest peak realized gains of the proposed cross bow-tie shaped antenna and the monopole antenna are about 4 dB and 2 dB, respectively. A prototype of the proposed dual-polarized omnidirectional antenna was fabricated to verify the simulated results, and the measured results are in a good agreement with the simulated ones.INDEX TERMS Cross bow-tie, dual-polarized omnidirectional antenna, monopole, WLAN, 5G.