Humans are capable of learning a new fine-grained concept with very little supervision, e.g., few exemplary images for a species of bird, yet our best deep learning systems need hundreds or thousands ...of labeled examples. In this paper, we try to reduce this gap by studying the fine-grained image recognition problem in a challenging few-shot learning setting, termed few-shot fine-grained recognition (FSFG). The task of FSFG requires the learning systems to build classifiers for the novel fine-grained categories from few examples (only one or less than five). To solve this problem, we propose an end-to-end trainable deep network, which is inspired by the state-of-the-art fine-grained recognition model and is tailored for the FSFG task. Specifically, our network consists of a bilinear feature learning module and a classifier mapping module: while the former encodes the discriminative information of an exemplar image into a feature vector, the latter maps the intermediate feature into the decision boundary of the novel category. The key novelty of our model is a "piecewise mappings" function in the classifier mapping module, which generates the decision boundary via learning a set of more attainable sub-classifiers in a more parameter-economic way. We learn the exemplar-to-classifier mapping based on an auxiliary dataset in a meta-learning fashion, which is expected to be able to generalize to novel categories. By conducting comprehensive experiments on three fine-grained datasets, we demonstrate that the proposed method achieves superior performance over the competing baselines.
The task of multi-label image recognition is to predict a set of object labels that present in an image. As objects normally co-occur in an image, it is desirable to model the label dependencies to ...improve the recognition performance. To capture and explore such important information, we propose graph convolutional networks (GCNs) based models for multi-label image recognition, where directed graphs are constructed over classes and information is propagated between classes to learn inter-dependent class-level representations. Following this idea, we design two particular models that approach multi-label classification from different views. In our first model, the prior knowledge about the class dependencies is integrated into classifier learning. Specifically, we propose Classifier Learning GCN (C-GCN) to map class-level semantic representations (e.g., word embeddings) into classifiers that maintain the inter-class topology. In our second model, we decompose the visual representation of an image into a set of label-aware features and propose prediction learning GCN (P-GCN) to encode such features into inter-dependent image-level prediction scores. Furthermore, we also present an effective correlation matrix construction approach to capture inter-class relationships and consequently guide information propagation among classes. Empirical results on generic multi-label image recognition demonstrate that both of the proposed models can obviously outperform other existing state-of-the-arts. Moreover, the proposed methods also show advantages in some other multi-label classification related applications.
Plants possess the remarkable ability to integrate the circadian clock with various signalling pathways, enabling them to quickly detect and react to both external and internal stress signals. ...However, the interplay between the circadian clock and biological processes in orchestrating responses to environmental stresses remains poorly understood. TOC1, a core component of the plant circadian clock, plays a vital role in maintaining circadian rhythmicity and participating in plant defences. Here, our study reveals a direct interaction between TOC1 and the promoter region of MYB44, a key gene involved in plant defence. TOC1 rhythmically represses MYB44 expression, thereby ensuring elevated MYB44 expression at dawn to help the plant in coping with lowest temperatures during diurnal cycles. Additionally, both TOC1 and MYB44 can be induced by cold stress in an Abscisic acid (ABA)‐dependent and independent manner. TOC1 demonstrates a rapid induction in response to lower temperatures compared to ABA treatment, suggesting timely flexible regulation of TOC1‐MYB44 regulatory module by the circadian clock in ensuring a proper response to diverse stresses and maintaining a balance between normal physiological processes and energy‐consuming stress responses. Our study elucidates the role of TOC1 in effectively modulating expression of MYB44, providing insights into the regulatory network connecting the circadian clock, ABA signalling, and stress‐responsive genes.
Summary statement
We present initial evidence that MYB44, a key player in the ABA signalling pathway, is directly targeted by the circadian clock. Additionally, our findings offer a comprehensive understanding of the TOC1‐MYB44 module, elucidating its crucial role under both ABA‐dependent and ABA‐independent conditions during cold stress.
Our work focuses on tackling large-scale fine-grained image retrieval as ranking the images depicting the concept of interests (i.e., the same sub-category labels) highest based on the fine-grained ...details in the query. It is desirable to alleviate the challenges of both fine-grained nature of small inter-class variations with large intra-class variations and explosive growth of fine-grained data for such a practical task. In this paper, we propose attribute-aware hashing networks with self-consistency for generating attribute-aware hash codes to not only make the retrieval process efficient, but also establish explicit correspondences between hash codes and visual attributes. Specifically, based on the captured visual representations by attention, we develop an encoder-decoder structure network of a reconstruction task to unsupervisedly distill high-level attribute-specific vectors from the appearance-specific visual representations without attribute annotations. Our models are also equipped with a feature decorrelation constraint upon these attribute vectors to strengthen their representative abilities. Then, driven by preserving original entities' similarity, the required hash codes can be generated from these attribute-specific vectors and thus become attribute-aware. Furthermore, to combat simplicity bias in deep hashing, we consider the model design from the perspective of the self-consistency principle and propose to further enhance models' self-consistency by equipping an additional image reconstruction path. Comprehensive quantitative experiments under diverse empirical settings on six fine-grained retrieval datasets and two generic retrieval datasets show the superiority of our models over competing methods. Moreover, qualitative results demonstrate that not only the obtained hash codes can strongly correspond to certain kinds of crucial properties of fine-grained objects, but also our self-consistency designs can effectively overcome simplicity bias in fine-grained hashing.
Alzheimer's disease (AD) is a complex disorder influenced by both genetic and environmental components and has become a major public health issue throughout the world. Oxidative stress and ...inflammation play important roles in the evolution of those major pathological symptoms. Jatrorrhizine (JAT), a main component of a traditional Chinese herbal, coptidis rhizome, has been shown to have neuroprotective effects and we previously showed that it is also able to clear oxygen free radicals and reduce inflammatory responses. In this study, we demonstrated that JAT administration could alleviate the learning and memory deficits in AD. Furthermore, we also found that JAT treatment reduced the levels of Aβ plaques in the cortex and hippocampus of APP/PS1 double-transgenic mice. Other studies suggest that there are gut microbiome alterations in AD. In order to explore the underlying mechanisms between gut microbiota and AD, DNA sequencing for 16s rDNA V3-V4 was performed in fecal samples from APP/PS1 transgenic mice and C57BL/6 wild-type (WT) mice. Our results indicated that APP/PS1 mice showed less Operational Taxonomic Units (OTUs) abundance in gut microbiota than WT mice and with different composition. Furthermore, JAT treatment enriched OTUs abundance and alpha diversity in APP/PS1 mice compared to WT mice. High dose of JAT treatment altered the abundance of some specific gut microbiota such as the most predominant phylum Firmicutes and Bacteroidetes in APP/PS1 mice. In conclusion, APP/PS1 mice display gut dysbiosis, and JAT treatment not only improved the memory deficits, but also regulated the abundance of the microbiota. This may provide a therapeutic way to balance the gut dysbiosis in AD patients.
Unsupervised learning technology has caught up with or even surpassed supervised learning technology in general object classification (GOC) and person re-identification (re-ID). However, it has been ...discovered that the unsupervised learning of fine-grained visual classification (FGVC) is more difficult than GOC and person re-ID. To bridge the gap between unsupervised and supervised learning for FGVC, we investigate the essential factors (including feature extraction, clustering, and contrastive learning) for the performance gap between supervised and unsupervised FGVC. Furthermore, we propose a simple, effective, and practical method, termed as unsupervised fine-grained clustering learning (UFCL), to alleviate this gap. Three key issues are concerned and improved: First, we introduce a robust and powerful backbone, ResNet50-IBN, which has the ability of domain adaptation when we transfer ImageNet pre-trained models to FGVC tasks. Next, we propose to introduce HDBSCAN rather than DBSCAN for clustering, which can generate better clusters for adjacent categories with fewer hyper-parameters. Finally, we propose a weighted feature agent and its update mechanism to perform contrastive learning by employing pseudo labels with unavoidable noise, which can enhance the optimization process of learning the network's parameters. The effectiveness of our UFCL was confirmed on the CUB-200-2011, Oxford-Flowers, Oxford-Pets, Stanford-Dogs, Stanford-Cars, and FGVC-Aircraft datasets. Under the unsupervised FGVC setting, we achieved state-of-the-art results and examined the primary factors and crucial parameters to offer practical guidance.
•Three factors are investigated for the gap between unsupervised and supervised FGVC.•A simple, effective, and practical method (UFCL) is proposed to alleviate the gap.•HDBSCAN is introduced to generate better clusters for adjacent categories.•A weighted updating strategy is proposed to update the network with noise labels.•The state-of-the-art results are achieved on six public FGVC datasets.