Cross-domain recommendation (CDR) has received attention to solve the cold-start and data sparsity problems. Existing methods mainly focus on the information about overlapping users or items, neglecting to effectively and efficiently utilize the information about nonoverlapping users or items in the source domain. Currently, graph prompt learning is proposed to bridge the gap between the pretrained tasks and downstream tasks, which can fully use the information of the source domain. However, existing graph prompt based CDR methods are rare and solely focus on the user-item interaction graphs without considering extra auxiliary information. Therefore, in this article, we construct knowledge graphs (KGs) as auxiliary information and propose a novel knowledge-aware graph prompt tuning for CDR (KGP-CDR) model. First, the KGs of the source and target domains are constructed, respectively, and a graph encoder is pretrained on the KG of the source domain. In addition, two types of graph prompts are designed: soft graph prompts and personalized graph prompts. These prompts are finetuned in the target domain along with the pretrained graph encoder. Ultimately, the predicted rating can be acquired through the finetuned graph prompts. Experiments on five real-world datasets show that the proposed method performs better than the state-of-the-art methods.

Over the past few years, cross-domain recommendation has gained great attention to resolve the cold-start issue. Many existing cross-domain recommendation methods model a preference bridge between the source and target domains to transfer preferences by the overlapping users. However, when there are insufficient cross-domain users available to bridge the two domains, it will negatively impact the recommender system's accuracy (ACC) and performance. Therefore, in this article, we propose to create a link between the source and the target domains by leveraging knowledge graph (KG) as the auxiliary information, and propose a novel knowledge-reinforced cross-domain recommendation (KR-CDR) method. First of all, we construct a new cross-domain KG (CDKG) by using the KGs that represent the source and target domains, respectively. Additionally, we employ reinforcement learning (RL) with meta learning on CDKG to discover meta-paths between the source and target domains. With these meta-paths, we obtain meta-path aggregated embedding vectors for cold-start users. Ultimately, the predicted rating can be acquired from the user meta-path aggregated embedding vector and item embedding vector. Experiments carried out on five real-world datasets show that the proposed method performs better than the state-of-the-art methods.

Abstract Cross‐domain object detection aims to generalize the distribution of features extracted by an object detector from an annotated domain to an unknown and unlabelled domain. Although one‐stage cross‐domain object detectors have significant advantages in deployment than two‐stage ones, they suffer from two problems. First, neglect of category features and inaccurate alignment between multiple category features would lead to decreased domain adaptation efficiency. Second, one‐stage detectors are more sensitive to imbalance of samples and negative samples severely affect the alignment process of domain adaptation. To overcome these two problems, an innovative category‐related attention domain adaptive method that refines discrimination for each category's feature has been proposed in this paper. In the proposed method, a group of domain discriminators is assigned to each category to refine the fine‐grained features between categories. The discriminators are trained via an adversarial discriminant framework to align the fine‐grained distributions cross different domains. A category attention alignment (CAA) module is proposed to navigate more attention to the foreground regions in instance‐level, which effectively alleviates the negative migration problem caused by the positive and negative sample imbalance of the one‐stage detector. Specifically, two sub‐modules in the CAA module are developed: a local CAA module and a global CAA module. These modules aim to optimize the domain offsets in both the local and global dimensions. In addition, a progressive global alignment module is designed to align image‐level features, offering prior knowledge of migration for the CAA module. The progressive global alignment module and CAA module collaboratively engage in benign competition with the backbone network across various levels. Extensive transferring experiments are conducted among cityscapes, foggy cityscapes, SIM10K, and KITTI. Experimental results show that the proposed method has much superior performance than other one‐stage cross‐domain detectors.

Graph Neural Networks (GNNs) based recommendations have shown significant performance improvement by explicitly modeling the user-item interactions as a bipartite graph. However, the existing GNNs-based recommendation methods suffer from the over-smoothing problem caused by utilizing the uniform distance of the reception field. To address this issue, we propose to explicitly incorporate the higher-order smoothness information into the node representation learning, and propose a new GNNs-based recommendation model namedHigher-orderSmoothness enhancedGraphCollaborativeFiltering (HS-GCF). The proposed model is mainly composed of two parts, namely lower-order module and higher-order module. The lower-order module guarantees that the lower-order smoothness is well obtained by using the user-item interactions. The higher-order module uses the latent group assumption to restrict too much noise introduced by the uniform distance property, which we call the higher-order smoothness information. Experiments are conducted on three real-world public datasets, and the experimental results show the performance improvements compared with several state-of-the-art methods and verify the importance of explicitly incorporating the higher-order smoothness information into the node representation learning.

Domain shift poses a significant challenge in Cross-Domain Facial Expression Recognition (CD-FER) due to the distribution variation between the source and target domains. Current algorithms mainly focus on learning domain-invariant features through global feature adaptation, while neglecting the transferability of local features across different domains. Additionally, these algorithms lack discriminative supervision during training on target datasets, resulting in deteriorated feature representation in the target domain. To address these limitations, we propose an Adaptive Global-Local Representation Learning and Selection (AGLRLS) framework. The framework incorporates global-local adversarial adaptation and semantic-aware pseudo label generation to enhance the learning of domain-invariant and discriminative feature representation during training. Meanwhile, a global-local prediction consistency learning is introduced to improve classification results during inference. Specifically, the framework consists of separate global-local adversarial learning modules that learn domain-invariant global and local features independently. We also design a semantic-aware pseudo label generation module, which computes semantic labels based on global and local features. Moreover, a novel dynamic threshold strategy is employed to learn the optimal thresholds by leveraging independent prediction of global and local features, ensuring filtering out the unreliable pseudo labels while retaining reliable ones. These labels are utilized for model optimization through the adversarial learning process in an end-to-end manner. During inference, a global-local prediction consistency module is developed to automatically learn an optimal result from multiple predictions. To validate the effectiveness of our framework, we conduct comprehensive experiments and analysis based on a fair evaluation benchmark. The results demonstrate that the proposed framework outperforms the current competing methods by a substantial margin.

Abstract Graph collaborative filtering methods have shown great performance improvements compared with deep neural network-based models. However, these methods suffer from data sparsity and data noise problems. To address these issues, we propose a new contrastive learning-based graph collaborative filtering method to learn more robust representations. The proposed method is called signal contrastive enhanced graph collaborative filtering (SC-GCF), which conducts contrastive learning on graph signals. It has been proved that graph neural networks correspond to low-pass filters on the graph signals from the graph convolution perspective. Different from the previous contrastive learning-based methods, we first pay attention to the diversity of graph signals to directly optimize the informativeness of the graph signals. We introduce a hypergraph module to strengthen the representation learning ability of graph neural networks. The hypergraph learning module utilizes a learnable hypergraph structure to model the latent global dependency relations that graph neural networks cannot depict. Experiments are conducted on four public datasets, and the results show significant improvements compared with the state-of-the-art methods, which confirms the importance of considering signal-level contrastive learning and hypergraph learning.

Anomaly detection on attributed networks has received an increasing amount of attention in recent years. Despite the success, most of the existing methods only focus on detecting the abnormal nodes while fail to detect the abnormal subgraphs. In this paper, we define a new problem of hybrid-order anomaly detection on attributed networks, which aims to detect both of the abnormal nodes and subgraphs. To this end, a new deep learning model called Hybrid-Order Graph Attention Network (HO-GAT) is developed, which is able to simultaneously detect the abnormal nodes and motif instances in an attributed network. In order to model the mutual influence between nodes and motif instances, the learning procedures of the node representation and the motif instance representation are integrated into a unified graph attention network with a novel hybrid-order self-attention mechanism. After learning the node representation and the motif instance representation, two decoders are respectively designed to reconstruct the attribute information of the nodes and motif instances, and the hybrid-order topological structure among nodes and motif instances. And finally, the reconstruction errors are utilized as the abnormal score of nodes and motif instances respectively. Extensive experiments conducted on real-world datasets have confirmed the effectiveness of the HO-GAT method.

Domain shift is one of the knotty problems that seriously restricts the accuracy of cross-domain facial expression recognition. Most existing works mainly focus on learning domain-invariant features by global feature adaption, and little works are conducted using the local features which are more transferable across different domains. In this paper, a consistent global-local feature and semantic learning framework is proposed which can learn domain-invariant global and local feature representation, and generate pseudo labels to facilitate cross-domain facial expression recognition. Specifically, the proposed method first simultaneously learns the domain-invariant global and local features via separately adversarial global and local learning. Once those features are acquired, a global and local semantic consistency is introduced to help generate pseudo labels for unlabeled data of the target dataset. By performing such strategy, more efficiency pseudo labels with high accuracy are produced due to the information diversity in global-local features and do without the image transformation. We conduct extensive experiments and analyses on several public datasets to demonstrate the effectiveness of the proposed model.

Model drifting is one of the knotty problems that seriously restricts the accuracy of discriminative trackers in visual tracking. Most existing works usually focus on improving the robustness of the target appearance model. However, they are prone to suffer from model drifting due to the inappropriate model updates during the tracking-by-detection. In this paper, we propose a novel update-pacing framework to suppress the occurrence of model drifting in visual tracking. Specifically, the proposed framework first initializes an ensemble of trackers, each of which updates the model in a different update interval. Once the forward tracking trajectory of each tracker is determined, the backward trajectory will also be generated by the current model to measure the difference with the forward one, and the tracker with the smallest deviation score will be selected as the most robust tracker for the remaining tracking. By performing such self-examination on trajectory pairs, the framework can effectively preserve the temporal context consistency of sequential frames to avoid learning corrupted information. To further improve the performance of the proposed method, a multi-feature extension framework is also proposed to incorporate multiple features into the ensemble of the trackers. The extensive experimental results obtained on large-scale object tracking benchmarks demonstrate that the proposed framework significantly increases the accuracy and robustness of the underlying base trackers, such as DSST, Struck, KCF, and CT, and achieves superior performance compared with the state-of-the-art methods without using deep models.

Object categories inherently form a hierarchy with different levels of concept abstraction, especially for fine-grained categories. For example, birds (Aves) can be categorized according to a four-level hierarchy of order, family, genus, and species. This hierarchy encodes rich correlations among various categories across different levels, which can effectively regularize the semantic space and thus make prediction less ambiguous. However, previous studies of fine-grained image recognition primarily focus on categories of one certain level and usually overlook this correlation information. In this work, we investigate simultaneously predicting categories of different levels in the hierarchy and integrating this structured correlation information into the deep neural network by developing a novel Hierarchical Semantic Embedding (HSE) framework. Specifically, the HSE framework sequentially predicts the category score vector of each level in the hierarchy, from highest to lowest. At each level, it incorporates the predicted score vector of the higher level as prior knowledge to learn finer-grained feature representation. During training, the predicted score vector of the higher level is also employed to regularize label prediction by using it as soft targets of corresponding sub-categories. To evaluate the proposed framework, we organize the 200 bird species of the Caltech-UCSD birds dataset with the four-level category hierarchy and construct a large-scale butterfly dataset that also covers four level categories. Extensive experiments on these two and the newly-released VegFru datasets demonstrate the superiority of our HSE framework over the baseline methods and existing competitors.

This paper proposes a novel framework to alleviate the model drift problem in visual tracking, which is based on paced updates and trajectory selection. Given a base tracker, an ensemble of trackers is generated, in which each tracker's update behavior will be paced and then traces the target object forward and backward to generate a pair of trajectories in an interval. Then, we implicitly perform self-examination based on trajectory pair of each tracker and select the most robust tracker. The proposed framework can effectively leverage temporal context of sequential frames and avoid to learn corrupted information. Extensive experiments on the standard benchmark suggest that the proposed framework achieves superior performance against state-of-the-art trackers.

In this paper, we propose a fast computational framework based on height functions descriptor for handling shape matching. To improve the efficiency, we utilize the strategy of heuristic search to reduce the large search space of dynamic programming (DP) algorithm between sample points of every two shapes during shape matching. Experiments on several public shape benchmarks(such as, MPEG-7 dataset, Kimia's dataset and ETH-80 dataset) demonstrate superior efficiency and competitive retrieval performance over previous methods.