Search

Domain adaptation and adversarial networks are two main approaches for transfer learning. Domain adaptation methods match the mean values of source and target domains, which requires a very large batch size during training. However, adversarial networks are usually unstable when training. In this paper, we propose a joint method of feature matching and adversarial networks to reduce domain discrepancy and mine domaininvariant features from the local and global aspects. At the same time, our method improves the stability of training. Moreover, the method is embedded into a unified convolutional neural network that can be easily optimized by gradient descent. Experimental results show that our joint method can yield the state-of-the-art results on three common public datasets.

The plant materials used in this subject were from two cultivars of pummelo (Citrus maxima) that were collected from Rong County, Yulin city, Guangxi province, China. One cultivar is a self-incompatible ‘Shatian’ pummelo. The other cultivar is the ‘Guiyou No.1’ pummelo that was derived from a bud sport variation of ‘Shatian’. Floral organs of ‘Shatian’ and ‘Guiyou No.1’ were collected, frozen immediately in liquid nitrogen, and then stored at -80°C for further analysis. Floral tissues were used for transcriptome sequencing (RNA-Seq).

The plant materials used in this subject were from two cultivars of pummelo (Citrus maxima) that were collected from Rong County, Yulin city, Guangxi province, China. One cultivar is a self-incompatible ‘Shatian’ pummelo. The other cultivar is the ‘Guiyou No.1’ pummelo that was derived from a bud sport variation of ‘Shatian’. Floral organs of ‘Shatian’ and ‘Guiyou No.1’ were collected, frozen immediately in liquid nitrogen, and then stored at -80°C for further analysis. Floral tissues were used for Whole-Genome bisulfite sequencing (WGBS).

The plant material used in this subject was from a cultivar of pummelo ("Guiyou No.1" pummelo, Citrus maxima) that was collected from Rong County, Yulin city, Guangxi province, China. Fresh leaf tissues were used for genomic sequencing. The sequencing of single-molecules was performed on the PacBio Sequel II platform (Pacific Biosciences, USA) by Beijing Biomarker Technologies Co. Ltd. (Beijing, China). A SMRTbell library with a size of 30 kb was constructed for genomic sequencing following the standard protocol provided by Pacific Biosciences. A total of 220.94 Gb CLR reads were assembled using Canu (v2.1.1) with the following parameters: -- pacbio (minReadLength = 2000, minOverlapLength = 500, corOutCoverage = 150, corMinCoverage =2 "batOptions = -dg 3 -db 3 -dr 1 -ca 500 -cp 50", corrected error rate= 0.035). Canu generated two assemblies composed of contigs and unitigs. The unitig assembly consisted of the contigs that split at any alternative path in the assembly graph. The contig assembly had longer continuity but more chimeric fragments as revealed by the genetic mapping analysis. The unitigs were polished iteratively using two rounds of Pilon with ~12 Gb of WGS BGISEQ data to generate the genome draft GYgv1.

Aimed at the necessary of digital video convert from RGB to BT.656 on many occasion,a realization scheme based on FPGA was put forward.The key technology and modules of system was analyzed.The status transfer mechanism was used to accomplish the production of BT.656 video timing;The linear and circular compensation method was applied to avoid the color distortion when conversion from RGB to YUV;The FIFO circuit with function of input and output selection control was employed to resolve the speed matching between input data and output data.The emulation result shows good purpose while an application example is introduced to indicate its broad prospect in video signal conversion.