Search

The dynamic expansion architecture is becoming popular in class incremental learning, mainly due to its advantages in alleviating catastrophic forgetting. However, task confusion is not well assessed within this framework, e.g., the discrepancy between classes of different tasks is not well learned (i.e., inter-task confusion, ITC), and certain priority is still given to the latest class batch (i.e., old-new confusion, ONC). We empirically validate the side effects of the two types of confusion. Meanwhile, a novel solution called Task Correlated Incremental Learning (TCIL) is proposed to encourage discriminative and fair feature utilization across tasks. TCIL performs a multi-level knowledge distillation to propagate knowledge learned from old tasks to the new one. It establishes information flow paths at both feature and logit levels, enabling the learning to be aware of old classes. Besides, attention mechanism and classifier re-scoring are applied to generate more fair classification scores. We conduct extensive experiments on CIFAR100 and ImageNet100 datasets. The results demonstrate that TCIL consistently achieves state-of-the-art accuracy. It mitigates both ITC and ONC, while showing advantages in battle with catastrophic forgetting even no rehearsal memory is reserved., Comment: 9 pages, 3 figures, Accepted on AAAI 2023

Background Osteoporosis (OP) is a systemic bone disease characterized by decreased bone mass, destruction of the bone tissue microstructure, increased bone brittleness and an increased risk of fracture. OP has a high incidence rate and long disease course and is associated with serious complications. Yigu decoction (YGD) is a compound prescription in traditional Chinese medicine that is used to treat OP. However, its mechanism in OP is not clear. This study used a tandem mass tag (TMT)quantitative proteomics method to explore the potential bone-protective mechanism of YGD in an osteoporotic rat model. Materials and methods A rat model of OP was established by ovariectomy. Eighteen 12-week-old specific-pathogen-free female Wistar rats weighing 220 ± 10 g were selected. The eighteen rats were randomly divided into 3 groups (n = 6 in each group): the normal, model and YGD groups. The right femurs from each group were subjected to quantitative biological analysis. TMT quantitative proteomics was used to analyze the proteins extracted from the bone tissue of rats in the model and YGD groups, and the differentially expressed proteins after intervention with YGD were identified as biologically relevant proteins of interest. Functional annotation correlation analysis was also performed to explore the biological function and mechanism of YGD. Result Compared with the model group, the YGD group showed significant upregulation of 26 proteins (FC > 1.2, P Conclusions YGD can significantly increase the bone mineral density (BMD) of osteoporotic rats and may play a therapeutic role by regulating target proteins involved in multiple signaling pathways. Therefore, these results improve the understanding of the OP mechanism and provide an experimental basis for the clinical application of YGD in OP treatment.

The Transformer-based encoder-decoder framework is becoming popular in scene text recognition, largely because it naturally integrates recognition clues from both visual and semantic domains. However, recent studies show that the two kinds of clues are not always well registered and therefore, feature and character might be misaligned in the difficult text (e.g., with rare shapes). As a result, constraints such as character position are introduced to alleviate this problem. Despite certain success, a content-free positional embedding hardly associates stably with meaningful local image regions. In this paper, we propose a novel module called Multi-Domain Character Distance Perception (MDCDP) to establish a visual and semantic related positional encoding. MDCDP uses positional embedding to query both visual and semantic features following the attention mechanism. The two kinds of constrained features are then fused to produce a reinforced feature, generating a content-aware embedding that well perceives spacing variations and semantic affinities among characters, i.e., multi-domain character distance. We develop a novel network named CDistNet that stacks multiple MDCDPs to guide a gradually precise distance modeling. Thus, the feature-character alignment is well built even various recognition difficulties presented. We create two series of augmented datasets with increasing recognition difficulties and apply CDistNet to both them and six public benchmarks. The experiments demonstrate that CDistNet outperforms recent popular methods by large margins in challenging recognition scenarios. It also achieves state-of-the-art accuracy on standard benchmarks. In addition, the visualization shows that CDistNet achieves proper information utilization in both visual and semantic domains. Our code is given in https://github.com/simplify23/CDistNet.

The plant height of rapeseed is one of the important factors that affects the production of rapeseed. If the plant height of rapeseed is too high, on the one hand, it will cause rapeseed to lodge and affect the yield, on the other hand, it will also affect the mechanized harvesting of rapeseed. In this research, the high-stalked line (YY50) and the dwarfed line (DW871) are crossed to obtain an F2 rapeseed population which was used to build pools, and then we used this to mine the main dwarfing genes. In the pools composed of tall and short stalks, we obtained 192.80Mb clean reads, which can be used for BSA (bulked segregant analysis). Preliminary positioning around the candidate section identified 23 SNP markers. Then 17 polymorphic SNP markers were obtained through polymorphism screening. Further we narrowed the candidate interval, and finally determined between 15.51-16.60Mb of ChrA10. Through identifying 231 genes from the above interval, it’s predicted that the production of dwarf traits may be related to lignin synthesis and limited inflorescence. It provides a basis for further mapping and cloning of the dwarfing gene DW871.