Language-Conditioned Manipulation (LCM) facilitates human-robot interaction via Behavioral Cloning (BC), which learns control policies from human demonstrations and serves as a cornerstone of embodied AI. Overcoming compounding errors in sequential action decisions remains a central challenge to improving BC performance. Existing approaches mitigate compounding errors through data augmentation, expressive representation, or temporal abstraction. However, they suffer from physical discontinuities and semantic-physical misalignment, leading to inaccurate action cloning and intermittent execution. In this paper, we present Continuous vision-language-action Co-Learning with Semantic-Physical Alignment (CCoL), a novel BC framework that ensures temporally consistent execution and fine-grained semantic grounding. It generates robust and smooth action execution trajectories through continuous co-learning across vision, language, and proprioceptive inputs (i.e., robot internal states). Meanwhile, we anchor language semantics to visuomotor representations by a bidirectional cross-attention to learn contextual information for action generation, successfully overcoming the problem of semantic-physical misalignment. Extensive experiments show that CCoL achieves an average 8.0% relative improvement across three simulation suites, with up to 19.2% relative gain in human-demonstrated bimanual insertion tasks. Real-world tests on a 7-DoF robot further confirm CCoL’s generalization under unseen and noisy object states.

Energy efficiency has become a critical requirement, driving recommendation systems for resource-constrained environments such as edge devices. Model quantization offers an effective way to build low-bitwidth models while preserving accuracy. However, user–item interaction graphs contain numerous nodes and complex topological structures, leading nodes to exhibit unique similarities and differences. Existing quantization methods uniformly process parameters in high-dimensional DNN layers (e.g., linear, convolutional, or attention layers), while inadequately capturing such similarities among node embeddings. This paper proposes GraphQ, a chunk-wise quantization framework for graph collaborative filtering that supports both the training and post-training phases in a unified perspective. Our core idea is to adaptively partition node embeddings into multiple chunks based on the distribution of embedding values, and then apply chunk-wise quantization. Specifically, for quantization-aware training (QAT), we introduce learnable low-precision quantization factors that partition node embeddings into multiple chunks and are dynamically updated following message passing. For post-training quantization (PTQ), we first cluster nodes and then partition their dimensions into chunks for weight clipping. Extensive experiments on four real-world datasets show that GraphQ outperforms state-of-the-art QAT methods by an average of 27.49% in Recall@10 under the 256-dimensional embedding and 2-bit settings, and surpasses PTQ methods by 78.64% on average under 4-bit settings.

202602 bcjz

Using low-rank adaptation to fine-tune pre-trained neural network models has attracted widespread attention due to its advantages of low resource requirements, high precision, and no additional inference delay. However, most existing methods are designed for large language models based on the Transformer structure and lack adaptation to convolutional neural networks (CNN) widely used on Internet of Things (IoT) devices. In addition, IoT devices are usually deployed outdoors and collect large amounts of data that are affected by environmental conditions. Providing a highly robust CNN model is a prerequisite for providing high-quality services. To this end, this paper proposes a highly robust seed-filter adaptation method (SeFA) for pre-trained CNNs. SeFA introduces an adaptation branch with the same structure as the backbone network. In the adaptation branch, some filters are first designated seed filters and grouped. Then, additional filters are generated from the grouped seed filters and nonlinear transformation functions (NLFs) with different hyperparameters. The seed filters' parameters are updated during model training, and the NLF hyperparameters are randomly initialized and frozen. Both grouping seed filters and configuring NLFs with non-learnable hyperparameters can improve the model's robustness. This is because grouping seed filters can generate diverse filters on demand without increasing the model's complexity, and the NLFs' rules can regularize the model. It is worth mentioning that the number of fine-tuning parameters of the pre-trained model that can adapt to downstream tasks can be flexibly controlled by specifying the number of seed filters. The key idea of this paper is to propose SeFA with flexible, controllable, learnable parameters, high robustness, and adaptability to pretrained CNN models, thereby facilitating fine-tuning on resource-constrained IoT devices and providing highly robust visual services. Experimental results on the CIFAR-10, CIFAR-10-C, CIFAR-100, CIFAR-100-C, and Icons-50 datasets demonstrate that the proposed SeFA outperforms other state-of-the-art methods. Specifically, on the ResNet-152 model and the CIFAR-10-C dataset, the accuracy of our SeFA is about 7% higher than that of the full fine-tuning method. In addition, we verify the efficiency of SeFA-based models for fine-tuning and inference on the device.

Cloud-assisted resource-constrained and widely used outdoor IoT devices deploying highly robust convolutional neural network (CNN) models to provide high-quality visual services have attracted widespread attention from industry and academia. Most existing methods suffer from two limitations: (i) large amount of parameter transmission, and (ii) low robustness in handling data affected by the environment. To this end, this paper proposes a transmission-friendly and robust CNN design method called NonLinear transformation generation Filter with NonLearnable hyperparameters, namely NL2Filter. In NL2Filter, some filters are first designated as seed filters, whose parameters are learnable, that is, updated as the model is trained. Other filters in this layer are generated based on the seed filters and nonlinear transformation function (NLF). The hyperparameters of the NLF in NL2Filter are randomly initialized and remain unchanged, so they can be saved and reproduced using random seed. After the cloud server trains the NL2Filter-CNN model, it only needs to send a small number of learnable parameters and random seed to reproduce the CNN model trained by the cloud server. Compared with sending the complete CNN model, sending only a small number of learnable parameters and random seed can significantly reduce the number of model parameters sent by the cloud server. On the other hand, NL2Filter groups seed filters and then uses NLFs with different hyperparameters to generate diverse filters on demand for each group, thereby improving the model's ability to capture feature diversity without increasing the complexity of the model, thereby improving the model's robustness in processing data affected by the environment. Experimental results on CIFAR-10, CIFAR-10-C, and Icons-50 datasets demonstrate that the proposed NL2Filter outperforms other state-of-the-art methods. Specifically, using the ResNet-101 architecture, on the CIFAR-10-C, NL2Filter's accuracy is about 3.9% higher than that of MonoCNN; On the Icons-50, NL2Filter's accuracy is about 2.7% higher than that of MonoCNN and about 2.3% higher than that of the standard ResNet-101.

Traffic prediction is pivotal in intelligent transportation systems. Existing works focus mainly on improving overall accuracy, overlooking a crucial problem of whether prediction results will lead to biased decisions by transportation authorities. In practice, the uneven deployment of traffic sensors in different urban areas produces imbalanced data, making the traffic prediction model fail in some urban areas and leading to unfair regional decision-making that eventually severely affects equity and quality of residents’ life. Existing fairness machine learning models struggle to maintain fair traffic prediction over prolonged periods. Although these models might achieve fairness at certain time slots, this static fairness will break down as traffic conditions change. To fill this research gap, we investigate prolonged fair traffic prediction, introducing two novel fairness metrics, i.e., region-based static fairness and sensor-based dynamic fairness, tailored to fairness fluctuations over time and across areas. An innovative prolonged fairness traffic prediction framework, namely FairTP, is then proposed. FairTP achieves prolonged fairness by alternating between “sacrifice” and “benefit” the prediction accuracy of each traffic sensor or area, ensuring that the number of these two actions are balanced over time. Specifically, FairTP incorporates a state identification module to discriminate whether the traffic sensors or areas are in a “sacrifice” or “benefit” state, thereby enabling prolonged fairness-aware traffic predictions. Additionally, we devise a state-guided balanced sampling strategy to select training examples to further enhance prediction fairness by mitigating the performance disparities among areas with uneven sensor distribution over time. Extensive experiments in two real-world datasets show that FairTP significantly improves prediction fairness without causing significant accuracy degradation.

Literature reviews play a crucial role in scientific research for understanding the current state of research, identifying gaps, and guiding future studies on specific topics. However, the process of conducting a comprehensive literature review is yet time-consuming. This paper proposes a novel framework, collaborative knowledge minigraph agents (CKMAs), to automate scholarly literature reviews. A novel prompt-based algorithm, the knowledge minigraph construction agent (KMCA), is designed to identify relations between concepts from academic literature and automatically constructs knowledge minigraphs. By leveraging the capabilities of large language models on constructed knowledge minigraphs, the multiple path summarization agent (MPSA) efficiently organizes concepts and relations from different viewpoints to generate literature review paragraphs. We evaluate CKMAs on three benchmark datasets. Experimental results show the effectiveness of the proposed method, further revealing promising applications of LLMs in scientific research.

Given a source, destination, and required keywords, the Optimal Route with Collective Spatial Keywords ( ORCSK ) query aims to find the shortest route covering all keywords. Existing Point of Interest (POI) candidate set-based and path expansion-based methods frequently produce inferior route quality or excessive time overhead, particularly under large-scale query keywords. To address this challenge, we introduce the DO framework, which pioneers the employ Deep Reinforcement Learning for the ORCSK. Specifically, DO first integrates the spatial index with the H2H index to generate and refine high-quality candidate sets. Subsequently, DO utilizes a Transformer-based model to determine the optimal route from the sets. To effectively combine spatial distance and POI attributes, we propose a novel dual-cross encoder architecture. Furthermore, leveraging this architecture, we introduce a multi-route generating strategy, exploiting parallel computing to enhance route quality. Our experiments on real-life road networks demonstrate superior route quality and response time compared to the state-of-the-art method, with an average improvement of 1-2 orders of magnitude in response time, and maintain high efficiency even under large-scale query keywords or dynamic POI attributes scenarios.

Predicting crimes in urban environments is crucial for public safety, yet existing prediction methods often struggle to align the knowledge across diverse cities that vary dramatically in data availability of specific crime types. We propose HYpernetwork-enhanced Spatial Temporal Learning (HYSTL), a framework that can effectively train a unified, stronger crime predictor without assuming identical crime types in different cities’ records. In HYSTL, instead of parameterising a dedicated predictor per crime type, a hypernetwork is designed to dynamically generate parameters for the prediction function conditioned on the crime type of interest. To bridge the semantic gap between different crime types, a structured crime knowledge graph is built, where the learned representations of crimes are used as the input to the hypernetwork to facilitate parameter generation. As such, when making predictions for each crime type, the predictor is additionally guided by its intricate association with other relevant crime types. Extensive experiments are performed on two cities with non-overlapping crime types, and the results demonstrate HYSTL outperforms state-of-the-art baselines. Our code is available at: https://github.com/fdnkar/HYSTL

Large language models (LLMs) fundamentally reshape learning and teaching models, shifting tutoring systems from supporting individual learning to facilitating collaborative learning (CL) like task-oriented group discussions. However, existing AI tutors struggle to guide CL, as they seldom model the interactions between AI tutors and students. Therefore, they cannot scaffold students to complete tasks collaboratively, which impairs learning outcomes and pedagogy adaptability. Additionally, existing AI tutors fail to make use of CL theories to generate instructive feedback, which leads to undesirable interactions such as over-instruction and limits students' autonomy. In this paper, we propose an LLM-based collaborative agent that innovatively leverages pedagogical strategies to sense discussion stages, detect learning issues, identify the timing of intervention, and generate instructive feedback. To develop the agent, we first design a prompting strategy based on a CL theory, that is, the Community of Inquiry, to cultivate the agent to understand the discussion status. Second, a multi-agent interaction framework is proposed to simulate the collaborative learning behavior between AI tutors and students. Meanwhile, a synthetic task-oriented group discussion dataset, namely CLTeach, is generated, which consists of 27k manually-verified multi-party dialogues with fine-grained annotations of instructive feedback and explanations. Lastly, we use CLTeach to fine-tune the LLM agent, ultimately enabling it to generate instructive feedback at the right time to support students in CL. Extensive experiments demonstrate that our agent achieves state-of-the-art performance in feedback generation and has the potential to mimic human teachers effectively.

Accurate spatial-temporal (ST) traffic prediction plays an essential role in intelligent transportation systems. Existing advanced traffic prediction methods typically utilize spatial-temporal graph neural networks (STGNNs) to capture the ST correlations and achieve excellent prediction performance. However, our experimental investigation reveals that existing static and dynamic graph-based STGNNs still incur excessive noise and redundancy, and fail to discover robust and reliable ST correlations in traffic networks. Moreover, most methods cannot explain the underlying reasons behind the ST correlations. To solve these problems, we propose a novel S patial- T emporal G raph M odeling framework via A daptive contrastive learning (ST-GMA). Firstly, we design a robust augmentation learning module to generate high-level and robust data augmentations via a self-supervised task for modeling reliable correlations. Then, we develop an adaptive contrastive learning module to update correlation graphs by effectively selecting positive and negative augmentations, reducing redundant calculations, and providing insights into the correlation changes. Finally, ST-GMA integrates the generated correlation graphs with ST convolution blocks to conduct traffic prediction tasks. Experimental results on five real-world datasets demonstrate that ST-GMA not only achieves significant prediction performance compared with state-of-the-art methods but also exhibits a new perspective on the interpretability of correlation changes.

As a distributed machine learning method, federated learning (FL) can collaboratively train a global model with multiple devices without sharing the original data, thus protecting certain privacy. However, due to the strong heterogeneity of edge nodes (ENs) participating in FL, the quality of data uploaded to the parameter server (PS) varies significantly. Without an appropriate incentive mechanism, low-quality contributors may receive disproportionately high rewards, while high-quality contributors may lack sufficient motivation, leading to inefficient participation and suboptimal global model performance. Consequently, it is critical to develop an effective incentive mechanism to promote fairness for the FL process. To address the issues of existing FL incentive mechanisms lacking privacy protection performance analysis, we propose a fairness-aware incentive mechanism for multi-server FL in edge-enabled wireless differential privacy (DP) networks. Specifically, the wireless channel noise is used to provide DP protection for the local model gradients uploaded by ENs. Next, the interaction between the PSs and ENs is modeled as a Stackelberg game. Furthermore, we solve the Stackelberg game process using backward induction and theoretically propose optimal strategies for both the PSs and ENs. Finally, extensive numerical simulations using real datasets demonstrate the superior performance of our theoretical analysis of the proposed scheme.

In multivariate time series (MTS) analysis, data loss is a critical issue that degrades analytical model performance and impairs downstream tasks such as structural health monitoring (SHM) and traffic flow monitoring. In real-world applications, MTS is usually collected by multiple types of sensors, making MTS and correlations between variates heterogeneous. However, existing MTS imputation methods overlook the heterogeneous correlations by manipulating heterogeneous MTS as a homogeneous entity, leading to inaccurate imputation results. Besides, correlations between different data types vary due to ever-changing environmental conditions, forming dynamic correlations in MTS. How to properly learn the hidden correlation from heterogeneous MTS for accurate data imputation remains unresolved. To solve the problem, we propose a multistage graph convolutional network with spatial attention (MSA-GCN). In the first stage, we decompose heterogeneous MTS into several clusters with homogeneous data collected from identical sensor types and learn intracluster correlations. Then, we devise a GCN with spatial attention to explore dynamic intercluster correlations, which is the second stage of MSA-GCN. In the last stage, we decode the learned features from previous stages via stacked convolutional neural networks. We jointly train these three-stage models to predict the missing data in MTS. Leveraging this multistage architecture and spatial attention mechanism makes MSA-GCN effectively learn heterogeneous and dynamic correlations among MTS, resulting in superior imputation performance. We tested MSA-GCN with the monitoring data from a large-span bridge and Wetterstation weather dataset. The results affirm its superiority over baseline models, demonstrating its enhanced accuracy in reducing imputation errors across diverse datasets.

Damage or relocation of network devices necessitates the replacement of devices. Additionally, manufacturers of devices may also change due to factors such as price, policies, and functional limitations. This brings great needs of converting existing configurations to fix new devices and ensuring consistent network policy behaviors, which is configuration translation. Manually translating requires network administrators to possess significant configuration expertise and to spend extreme time translating thousands of commands for a core device. Therefore, we propose ConfigTrans, a novel configuration translation method for commands with and without parameters based on constraint solving and large language models (LLMs), respectively. For commands with parameters, we explore constraints about parameters, commands, and format limitations, and design a heuristic algorithm to solve them. For commands without parameters, we utilize an LLM to find the translation. The most likely several commands are selected based on semantic similarity, and the final result and keywords in it are chosen by the LLM. To meet the requirements of the view, the result commands are arranged while supplementing missing higherview commands. Experimental results show that our method achieves an accuracy rate of 82.47 % for translating commands with parameters and 75.5 % for commands without parameters.

Internet of Things (IoT) data provides rich data sources and application scenarios for trajectory representation learning. Trajectory representation learning aims to transform the original trajectory information into a general low-dimensional vector representation for many different downstream tasks (trajectory similarity calculation, anomaly detection, etc.). Current road network-based trajectory learning methods mainly focus on the spatial structure of the road network and often ignore the semantic information and complex feature information embedded in IOT data, in addition, the spatial and semantic properties of trajectories cannot be adequately preserved simultaneously. To this end, we propose a Trajectory Representation Learning framework based on Road network-TRLR. It first uses a graph attention network to learn the topological and semantic properties of road network segments, then it fuses node vectors and segment vectors as representations of node units, this is to support more data input types and enhanced node characteristics. Lastly, it learns the travel semantics of the trajectories through an information-enhanced transformer model, which captures the sequence information in the trajectory and generates the trajectory representation vector. In addition, we also propose four data augmentation methods to ensure that the trajectories can maintain both their spatial and semantic properties. To validate the effectiveness of our modeling approach, we conduct experiments on real datasets for similar trajectory search and mask prediction tasks. The experimental results demonstrate the performance improvement of our model.

Recent studies successfully learned static graph embeddings that are structurally fair by preventing the effectiveness disparity of high- and low-degree vertex groups in downstream graph mining tasks. However, achieving structure fairness in dynamic graph embedding remains an open problem. Neglecting degree changes in dynamic graphs will significantly impair embedding effectiveness without notably improving structure fairness. This is because the embedding performance of high-degree and low-to-high-degree vertices will significantly drop close to the generally poorer embedding performance of most slightly changed vertices in the long-tail part of the power-law distribution. We first identify biased structural evolutions in a dynamic graph based on the evolving trend of vertex degree and then propose FairDGE, the first structurally Fair Dynamic Graph Embedding algorithm. FairDGE learns biased structural evolutions by jointly embedding the connection changes among vertices and the long-short-term evolutionary trend of vertex degrees. Furthermore, a novel dual debiasing approach is devised to encode fair embeddings contrastively, customizing debiasing strategies for different biased structural evolutions. This innovative debiasing strategy breaks the effectiveness bottleneck of embeddings without notable fairness loss. Extensive experiments demonstrate that FairDGE achieves simultaneous improvement in the effectiveness and fairness of embeddings.

Multimodal knowledge graph (MKG) reasoning has attracted significant attention since impressive performance has been achieved by adding multimodal auxiliary information (i.e., texts and images) to the entities of traditional KGs. However, existing studies heavily rely on path-based methods for learning structural modality, failing to capture the complex structural interactions among multimodal entities beyond the reasoning path. In addition, existing studies have largely ignored the dynamic impact of different multimodal features on different decision facts for reasoning, which utilize asymmetric coattention to independently learn the static interplay between different modalities without dynamically joining the reasoning process. We propose a novel Dynamic Structure-aware representation learning method, namely DySarl, to overcome this problem and significantly improve the MKG reasoning performance. Specifically, we devise a dual-space multihop structural learning module in DySarl, aggregating the multihop structural features of multimodal entities via a novel message-passing mechanism. It integrates the message paradigms in Euclidean and hyperbolic spaces, effectively preserving the neighborhood information beyond the limited multimodal query paths. Furthermore, DySarl has an interactive symmetric attention module to explicitly learn the dynamic impacts of unimodal attention senders and multimodal attention targets on decision facts through a newly designed symmetric attention component and fact-specific gated attention unit, equipping DySarl with the dynamic associations between the multimodal feature learning and later reasoning. Extensive experiments show that DySarl achieves significantly improved reasoning performance on two public MKG datasets compared with that of the state-of-the-art baselines. Source codes are available at https://github.com/HUSTNLP-codes/DySarl.

Personality Recognition in Conversation (PRC) aims to identify the personality traits of speakers through textual dialogue content. It is essential for providing personalized services in various applications of Human-Computer Interaction (HCI), such as AI-based mental therapy and companion robots for the elderly. Most recent studies analyze the dialog content for personality classification yet overlook two major concerns that hinder their performance. First, crucial implicit factors contained in conversation, such as emotions that reflect the speakers' personalities are ignored. Second, only focusing on the input dialog content disregards the semantic understanding of personality itself, which reduces the interpretability of the results. In this paper, we propose Affective Natural Language Inference (Affective-NLI) for accurate and interpretable PRC. To utilize affectivity within dialog content for accurate person-ality recognition, we fine-tuned a pre-trained language model specifically for emotion recognition in conversations, facilitating real-time affective annotations for utterances. For interpretability of recognition results, we formulate personality recognition as an NLI problem by determining whether the textual description of personality labels is entailed by the dialog content. Extensive experiments on two daily conversation datasets suggest that Affective-NLI significantly outperforms (by 6%-7%) state-of-the-art approaches. Additionally, our Flow experiment demonstrates that Affective-NLI can accurately recognize the speaker's personality in the early stages of conversations by surpassing state-of-the-art methods with 22% −34%11Our source code and data is at https://github.com/preke/Affective-NLI..

Computation offloading involves transmitting compute-intensive tasks from mobile users (MUs) to edge environments. This compensates for the limitations of terminal devices in computing performance and resource storage capabilities. However, with the growing demand for offloading compute-intensive tasks, the burden on edge server networks has intensified. On the other hand, the response time of application has increased, leading to a decline in user experience. Therefore, selecting a reliability metric caused by congestion becomes an important issue for analyzing the quantitative characteristics of edge networks. The current research primarily focuses on the precise validation offloading task results. Typically, methods like task redistribution and third-party assistance are employed to ensure task reliability. However, these measures lead to increased time and energy consumption. In this paper, we propose a more versatile reliability metric based on the probability distribution of average waiting time in a multi-queue model. Additionally, to incentive MUs to offload more tasks and enhance the economic utility of edge server providers (ESPs), we employ a Stackelberg game to model the dynamic interaction between ESPs and MUs. Finally, we utilize the alternating direction method of multipliers (ADMM) algorithm to derive the optimal strategies for ESPs and MUs. Simulation results demonstrate that our proposed approach surpasses the baselines in terms of reliability indicator. Moreover, it achieves faster convergence and decision-making in comparison to conventional heuristic methods.

Accurate prediction of ocean factors (e.g., temperature and salinity) is crucial for plenty of applications, including weather forecasting, storm tracking, and ecosystem protection. Meanwhile, it is well-known that the ocean is a unified system and various ocean factors usually influence each other. For example, the changes in temperature would affect the distribution of salinity in ocean. However, the existing studies for ocean factor prediction mainly focus on designing individual models for predicting specific factors and ignore the correlations between different factors, thus having potentials to be further improved. Therefore, we propose a unified framework UniOcean to predict multiple ocean factors simultaneously, and capture the correlations between them to improve the prediction accuracy. First, considering that ocean factors are usually collected with different temporal scales, we develop the fine-grained multiscale data fusion module to integrate multiple ocean factors with different temporal scales, and effectively learn their hierarchical patterns at different levels. Then, since the correlations between ocean factors may vary across different time periods, the multifactor correlation learning module is constructed to adaptively learn the dynamic correlations between different factors. Finally, we utilize the factor-specific towers to predict multiple ocean factors simultaneously. Experimental results on five real-world remote sensing datasets demonstrate that UniOcean significantly improves the prediction accuracy by 11%–53% in terms of MSD for different ocean factors.

Compared with only pursuing recommendation accuracy, the explainability of a recommendation model has drawn more attention in recent years. Many graph-based recommendations resort to informative paths with the attention mechanism for the explanation. Unfortunately, these attention weights are intentionally designed for model accuracy but not explainability. Recently, some researchers have started to question attention-based explainability because the attention weights are unstable for different reproductions, and they may not always align with human intuition. Inspired by the counterfactual reasoning from causality learning theory, we propose a novel explainable framework targeting path-based recommendations, wherein the explainable weights of paths are learned to replace attention weights. Specifically, we design two counterfactual reasoning algorithms from both path representation and path topological structure perspectives. Moreover, unlike traditional case studies, we also propose a package of explainability evaluation solutions with both qualitative and quantitative methods. We conduct extensive experiments on four real-world datasets, the results of which further demonstrate the effectiveness and reliability of our method.

POI recommendation is practically important to facilitate various Location-Based Social Network (LBSN) services, and has attracted rising research attention recently. Existing works generally assume the available POI check-ins reported by users are the ground-truth depiction of user behaviors. However, in real application scenarios, the check-in data can be rather unreliable (e.g. sparse, incomplete and inaccurate) due to both subjective and objective causes including positioning error and user privacy concerns. The data uncertainty issue may lead to significant negative impacts on the performance of the POI recommendation, but is not fully explored by existing works. To this end, this paper investigates a novel problem of robust POI recommendation by considering the uncertainty factors of the user check-ins, and proposes a Bayes-enhanced Multi-view Attention Network (BayMAN for short) to effectively address it. Specifically, we construct three POI graphs to comprehensively model the dependencies among the POIs from different views, including the personal POI transition graph, the semantic-based POI graph and distance-based POI graph. As the personal POI transition graph is usually sparse and sensitive to noise, we design a Bayes-enhanced spatial dependency learning module for data augmentation from the local view. A Bayesian posterior guided graph augmentation approach is adopted to generate a new graph with collaborative signals to increase the data diversity. Then both the original and the augmented graphs are used for POI representation learning to counteract the data uncertainty issue. Next, the POI representations of the three view graphs are input into the proposed multi-view attention-based user preference learning module. By incorporating the semantic and distance correlations of POIs, the user preference can be effectively refined and finally robust recommendation results are achieved. We conduct extensive experiments over three real-world LSBN datasets. The results show that BayMAN significantly outperforms the state-of-the-art methods in POI recommendation when the available check-ins are incomplete and noisy.

Inferring the fine-grained urban flows based on the coarse-grained flow observations is practically important to many smart city-related applications. However, the collected human/vehicle trajectory flows are usually rather unreliable, may contain various noise and sometimes are incomplete, thus posing great challenges to existing approaches. In this paper, we present a pioneering study on robust fine-grained urban flow inference with noisy and incomplete urban flow observations, and propose a denoising diffusion model named DiffUFlow to effectively address it. Specifically, we propose an improved reverse diffusion strategy. A spatial-temporal feature extraction network called STFormer and a semantic features extraction network called ELFetcher are also proposed. Then, we overlay the spatial-temporal feature map extracted by STFormer onto the coarse-grained flow map, serving as a conditional guidance for the reverse diffusion process. We further integrate the semantic features extracted by ELFetcher to cross-attention layers, enabling the comprehensive consideration of semantic information encompassing the entirety of urban data in fine-grained inference. Extensive experiments on two large real-world datasets validate the effectiveness of our method compared with the state-of-the-art baselines.

Dynamic graph embedding learns representation vectors for vertices and edges in a graph that evolves over time. We aim to capture and embed the evolution of vertices' temporal connectivity. Existing work studies the vertices' dynamic connection changes but neglects the time it takes for edges to evolve, failing to embed temporal linkage information into the evolution of the graph. To capture vertices' temporal linkage evolution, we model dynamic graphs as a sequence of snapshot graphs, appending the respective timespans of edges (ToE). We co-train a linear regressor to embed ToE while inferring a common latent space for all snapshot graphs by a matrix-factorization-based model to embed vertices' dynamic connection changes. Vertices' temporal linkage evolution is captured as their moving trajectories within the common latent representation space. Our embedding algorithm converges quickly with our proposed training methods, which is very time efficient and scalable. Extensive evaluations on several datasets show that our model can achieve significant performance improvements, i.e. 22.98% on average across all datasets, over the state-of-the-art baselines in the tasks of vertex classification, static and time-aware link prediction, and ToE prediction.

Dynamic graphs refer to graphs whose structure dynamically changes over time. Despite the benefits of learning vertex representations (i.e., embeddings) for dynamic graphs, existing works merely view a dynamic graph as a sequence of changes within the vertex connections, neglecting the crucial asynchronous nature of such dynamics where the evolution of each local structure starts at different times and lasts for various durations. To maintain asynchronous structural evolutions within the graph, we innovatively formulate dynamic graphs as temporal edge sequences associated with joining time of vertices (ToV) and timespan of edges (ToE). Then, a time-aware Transformer is proposed to embed vertices’ dynamic connections and ToEs into the learned vertex representations. Meanwhile, we treat each edge sequence as a whole and embed its ToV of the first vertex to further encode the time-sensitive information. Extensive evaluations on several datasets show that our approach outperforms the state-of-the-art in a wide range of graph mining tasks. At the same time, it is very efficient and scalable for embedding large-scale dynamic graphs.

Balanced rule-constrained resource allocation aims to evenly distribute tasks to different processors under allocation rule constraints. Conventional heuristic approach fails to achieve optimal solution while simple brute force method has the defect of high computational complexity. To address these limitations, we propose recursive balanced k-subset sum partition (RBkSP), in which iterative 'cut-one-out' policy is employed that in each round, only one subset whose weight of tasks sums up to 1/k of the total weight of all tasks is taken out from the set. In a single partition, we first create a dynamic programming table with its elements recursively computed, then use 'zig-zag search' method to explore the table, find out elements with optimal subset partition and assign different partitions to proper places. Next, to resolve conflicts during allocation, we use simple but effective heuristic method to adjust the allocation of tasks that is contradicted to allocation rules. Testing results show RBkSP can achieve more balanced results with lower computational complexity over classical benchmarks.