Ultra-Fine-Grained Visual Categorization (ultra-FGVC) has become a popular problem due to its great real-world potential for classifying the same or closely related species with very similar layouts. However, there present many challenges for the existing ultra-FGVC methods, firstly there are always not enough samples in the existing ultraFGVC datasets based on which the models can easily get overfitting. Secondly, in practice, we are likely to find new species that we have not seen before and need to add them to existing models, which is known as incremental learning. The existing methods solve these problems by Few-Shot Class Incremental Learning (FSCIL), but the main challenge of the FSCIL models on ultra-FGVC tasks lies in their inferior discrimination detection ability since they usually use low-capacity networks to extract features, which leads to insufficient discriminative details extraction from ultrafine-grained images. In this paper, a self-supervised feature enhancement for the few-shot incremental learning network (SSFE-Net) is proposed to solve this problem. Specifically, a self-supervised learning (SSL) and knowledge distillation (KD) framework is developed to enhance the feature extraction of the low-capacity backbone network for ultra-FGVC few-shot class incremental learning tasks. Besides, we for the first time create a series of benchmarks for FSCIL tasks on two public ultra-FGVC datasets and three normal finegrained datasets, which will facilitate the development of the Ultra-FGVC community. Extensive experimental results on public ultra-FGVC datasets and other state-of-the-art benchmarks consistently demonstrate the effectiveness of the proposed method.

Few-shot class-incremental learning (FSCIL) aims to continually learn novel data with limited samples. One of the major challenges is the catastrophic forgetting problem of old knowledge while training the model on new data. To alleviate this problem, recent state-of-the-art methods adopt a well-trained static network with fixed parameters at incremental learning stages to maintain old knowledge. These methods suffer from the poor adaptation of the old model with new knowledge. In this work, a dynamic clustering and recovering network (DyCR) is proposed to tackle the adaptation problem and effectively mitigate the forgetting phenomena on FSCIL tasks. Unlike static FSCIL methods, the proposed DyCR network is dynamic and trainable during the incremental learning stages, which makes the network capable of learning new features and better adapting to novel data. To address the forgetting problem and improve the model performance, a novel orthogonal decomposition mechanism is developed to split the feature embeddings into context and category information. The context part is preserved and utilized to recover old class features in future incremental learning stages, which can mitigate the forgetting problem with a much smaller size of data than saving the raw exemplars. The category part is used to optimize the feature embedding space by moving different classes of samples far apart and squeezing the sample distances within the same classes during the training stage. Experiments show that the DyCR network outperforms existing methods on four benchmark datasets. The code is available at: https://github.com/zichengpan/DyCR.

We report ab initio total-energy calculations that examine the cubic-graphitic transformation and relative stability of two high density phases of cubic ${\mathrm{BC}}_2\mathrm{N}$ that have nearly identical crystal structure, lattice parameters, energetics and lattice dynamics, and equation of states over a large range of hydrostatic pressure. The calculated energy barriers for the formation and breaking of the bonds between the graphitic layers along different crystallographic directions clearly distinguish these two phases and provide important insights for the structural assignment of the experimentally synthesized high density cubic ${\mathrm{BC}}_2\mathrm{N}$. We also identify a graphitic structure as a possible precursor for the synthesis of the high density cubic phase that has the highest critical tensile strength among the ${\mathrm{BC}}_2\mathrm{N}$ polymorphs.

Modern inverter-based air-conditioners (IACs) are generally equipped with power-electronic components to achieve higher efficiency and more advanced controllability. Thus, their operating behaviours can be very different from the conventional air-conditioners with direct motor connections, and their new characteristics have not been adequately studied in the current literature. Thus, hardware experiments are conducted to extract the novel dynamic frequency response features of modern inverter-based air-conditioners under a range of different frequency disturbances. Then, a new IAC model with physical meanings is developed by accommodating the discovered new features such as time delays, load frequency relief and the minimum operating power limit. Further, this letter also demonstrates that the developed air-conditioner model can accurately represent the dynamics of the IACs under various frequency events. The extracted dynamic frequency response behaviours and the developed air-conditioner model will provide a solid foundation for the modern appliance modelling in a large scale and will assist to improve the accuracy of the inertia response representations of distribution network’s loads.