Federated learning (FL) is a hot collaborative training framework via aggregating model parameters of decentralized local clients. However, most FL methods unreasonably assume data categories are known and fixed in advance. Moreover, some new clients that collect novel unseen by other may be introduced to irregularly. These issues render global undergo catastrophic forgetting on old categories, when receive consecutively under limited memory storing categories. To tackle the above issues, we...

Incremental Named Entity Recognition (INER) involves the sequential learning of new entity types without accessing training data previously learned types. However, INER faces challenge catastrophic forgetting specific for incremental learning, further aggravated by background shift (i.e., old and future are labeled as non-entity type in current task). To address these challenges, we propose a method called task Relation Distillation Prototypical pseudo label (RDP) INER. Specifically, to...

Few-shot medical image segmentation has achieved great progress in improving accuracy and efficiency of analysis the biomedical imaging field. However, most existing methods cannot explore inter-class relations among base novel classes to reason unseen classes. Moreover, same kind class large intra-class variations brought by diverse appearances, shapes scales, thus causing ambiguous visual characterization degrade generalization performance these on To address above challenges, this paper,...

Abstract Current codes for fire protection of buildings are mainly based on the movement adults and neglect characteristic pre-school children. Having a profound comprehension difference between children passing bottlenecks is great help to improve safety levels preschool This paper presents an experimental study bottleneck flow in room. The characteristics children’s adults’ investigated with two macroscopic properties: density speed profiles as well microscopic time: motion activation...

This study investigated the influence of a cyclist on crowd behaviors shared road. A series mixed traffic experiments and pedestrians were conducted in circular wide track under different densities moving directions. The fundamental diagram showed clear difference unidirectional bidirectional scenarios. restricted pedestrian flow evidently flow, while critical density around 0.9 ped m−2 was observed, where restriction motion not noticeable unless reached. speed also depended relative...

Pedestrian dynamics of preschool children in facilities has become a rapidly growing research topic due to its significant implications for children's safety. This work investigates the movement characteristics 40 passing through 1.0 m wide right-angled corridor under low and high motivations. Under motivation, were asked walk as normal run quickly if an emergency, respectively. The fundamental diagrams, velocity directions, positions, spatial distributions are studied. Individual speed...

Unsupervised Domain Adaptation (UDA), which aims to explore the transferrable features from a well-labeled source domain related unlabeled target domain, has been widely progressed. Nevertheless, as one of mainstream, existing adversarial-based methods neglect filter irrelevant semantic knowledge, hindering adaptation performance improvement. Besides, they require an additional discriminator that strives extractor generate confused representations, but discrete designing may cause model...

The elderly ratio in the world's population is increasing and countries are facing ongoing challenges improving safety of pedestrians. It vital importance to organize evacuation pedestrians effectively under emergencies. In this study, simulation parameters Pathfinder GCFM (generalized centrifugal-force model) set according results control experiment. Through comparison, it can be found that more consistent with experimental results. As increases, time required get longer. sub room affected...

Understanding the movement characteristics of pedestrians is essential for management mass gatherings. In this work, a series experiments were conducted to study individuals through angled corridors with sharp-turn and curved-turn sections. The influence different angled-corridors on offset trajectory, speeds, turning point studied. An inward around sections observed regardless speeds directions. It also found that distance points (the relative difference between selected reference points)...

Pedestrian movement through bottlenecks have been widely studied from various aspects to understand the effects of on pedestrian flow. However, few attentions paid characteristics preschool children, who show obvious differences behaviour compared adults due poor balance and understanding danger especial under emergencies. In this study, we focus evacuation children with laboratory experiments. From all experiment, do not observe clear lane formation process trajectories diagrams. It is...

Deep unfolding methods have made impressive progress in restoring 3D hyperspectral images (HSIs) from 2D measurements through convolution neural networks or Transformers spectral compressive imaging. However, they cannot efficiently capture long-range dependencies using global receptive fields, which significantly limits their performance HSI reconstruction. Moreover, these may suffer local context neglect if we directly utilize Mamba to unfold a feature map as 1D sequence for modeling...

Few-shot medical image segmentation has achieved great progress in improving accuracy and efficiency of analysis the biomedical imaging field. However, most existing methods cannot explore inter-class relations among base novel classes to reason unseen classes. Moreover, same kind class large intra-class variations brought by diverse appearances, shapes scales, thus causing ambiguous visual characterization degrade generalization performance these on To address above challenges, this paper,...

Abstract The influence of different motivations on pedestrian evacuation efficiency, like whether incentivizing faster moving or encouraging polite behaviors is beneficial to evacuation, and the potential existence gender-based differences, still lacks clear answers. This study aims narrow this gap by conducting a laboratory bottleneck experiment investigate movement motivation gender pedestrians efficiency. Our findings reveal that both width significantly impact flow. For men, when...

Custom diffusion models (CDMs) have attracted widespread attention due to their astonishing generative ability for personalized concepts. However, most existing CDMs unreasonably assume that concepts are fixed and cannot change over time. Moreover, they heavily suffer from catastrophic forgetting concept neglect on old when continually learning a series of new To address these challenges, we propose novel Concept-Incremental text-to-image Diffusion Model (CIDM), which can resolve learn...