A5024168844- Advanced Graph Neural Networks
- Topic Modeling
- Recommender Systems and Techniques
- Semantic Web and Ontologies
- Natural Language Processing Techniques
- Robotic Path Planning Algorithms
- Robotics and Sensor-Based Localization
- Graph Theory and Algorithms
- UAV Applications and Optimization
- Bacterial Infections and Vaccines
- Multimedia Communication and Technology
- Advanced Frequency and Time Standards
- Service-Oriented Architecture and Web Services
- Caching and Content Delivery
- Advanced Bandit Algorithms Research
- Currency Recognition and Detection
- Insect symbiosis and bacterial influences
- Peer-to-Peer Network Technologies
- Geographic Information Systems Studies
- Teaching and Learning Programming
- Microbial Fuel Cells and Bioremediation
- Geophysics and Sensor Technology
- Mobile Agent-Based Network Management
- Stock Market Forecasting Methods
- Inertial Sensor and Navigation
University of Hong Kong
2022-2024
Sichuan Center for Disease Control and Prevention
2024
Sichuan Agricultural University
2024
Jinan University
2024
Northwestern Polytechnical University
2024
Chinese University of Hong Kong
2023
China Electronics Technology Group Corporation
2022
Institute of Electronics
2022
Nanjing Institute of Technology
2021
Wuhan University
2021
Knowledge Graphs (KGs) have been utilized as useful side information to improve recommendation quality. In those recommender systems, knowledge graph often contains fruitful facts and inherent semantic relatedness among items. However, the success of such methods relies on high quality graphs, may not learn representations with two challenges: i) The long-tail distribution entities results in sparse supervision signals for KG-enhanced item representation; ii) Real-world graphs are noisy...
Learning dynamic user preference has become an increasingly important component for many online platforms (e.g., video-sharing sites, e-commerce systems) to make sequential recommendations. Previous works have made efforts model item-item transitions over interaction sequences, based on various architectures, e.g., recurrent neural networks and self-attention mechanism. Recently emerged graph also serve as useful backbone models capture item dependencies in recommendation scenarios. Despite...
Current sequential recommender systems are proposed to tackle the dynamic user preference learning with various neural techniques, such as Transformer and Graph Neural Networks (GNNs). However, inference from highly sparse behavior data may hinder representation ability of pattern encoding. To address label shortage issue, contrastive (CL) methods recently perform augmentation in two fashions: (i) randomly corrupting sequence (e.g. stochastic masking, reordering); (ii) aligning...
In this paper, we introduce a new self-supervised rationalization method, called KGRec, for knowledge-aware recommender systems. To effectively identify informative knowledge connections, propose an attentive mechanism that generates rational scores triplets. With these scores, KGRec integrates generative and contrastive tasks recommendation through masking. highlight rationales in the graph, design novel task form of masking-reconstructing. By masking important with high is trained to...
Knowledge Graphs (KGs) have emerged as invaluable resources for enriching recommendation systems by providing a wealth of factual information and capturing semantic relationships among items. Leveraging KGs can significantly enhance performance. However, not all relations within KG are equally relevant or beneficial the target task. In fact, certain item-entity connections may introduce noise lack informative value, thus potentially misleading our understanding user preferences. To bridge...
Self-supervised learning (SSL) has gained significant interest in recent years as a solution to address the challenges posed by sparse and noisy data recommender systems. Despite growing number of SSL algorithms designed provide state-of-the-art performance various recommendation scenarios (e.g., graph collaborative filtering, sequential recommendation, social KG-enhanced recommendation), there is still lack unified frameworks that integrate across different domains. Such framework could...
Recently, exploiting a knowledge graph (KG) to enrich the semantic representation of news article have been proven be effective for recommendation. These solutions focus on learning articles with additional information in graph, where user representations are mainly derived based these later. However, different users would hold interests same article. In other words, directly identifying entities relevant user's interest and deriving resultant could enable better recommendation explanation.
GNN-based recommendation systems have been successful in capturing complex user-item interactions using multi-hop message passing. However, these methods often struggle to handle the dynamic nature of interactions, making it challenging adapt changes user preferences and new data distributions. This limits their scalability performance real-world scenarios. In our study, we propose a framework called GraphPro that combines graph pre-training with prompt learning an efficient way. unique...
Heterogeneous graph learning aims to capture complex relationships and diverse relational semantics among entities in a heterogeneous obtain meaningful representations for nodes edges. Recent advancements neural networks (HGNNs) have achieved state-of-the-art performance by considering relation heterogeneity using specialized message functions aggregation rules. However, existing frameworks limitations generalizing across datasets. Most of these follow the "pre-train" "fine-tune" paradigm on...
Graph Neural Networks (GNNs) have advanced graph structure understanding via recursive information exchange and aggregation among nodes. To improve model robustness, self-supervised learning (SSL) has emerged as a promising approach for data augmentation. However, existing methods generating pre-trained embeddings often rely on fine-tuning with specific downstream task labels, which limits their usability in scenarios where labeled is scarce or unavailable. address this, our research focuses...
Learning effective geospatial embeddings is crucial for a series of applications such as city analytics and earth monitoring. However, learning comprehensive region representations presents two significant challenges: first, the deficiency intra-region feature representation; second, difficulty from intricate inter-region dependencies. In this paper, we present GeoHG, an heterogeneous graph structure various downstream tasks. Specifically, tailor satellite image representation through...
Abstract Unmanned aerial vehicle (UAV) detection has the advantages of flexible deployment and no casualties. It become a force that cannot be ignored in battlefield. Scientific efficient mission planning can help improving survival rate completion UAV search dynamic environments. Towards problem collaborative for multi-types time-sensitive moving targets, algorithm based on hybrid layered artificial potential fields (HL-APF) was proposed. This method consists two parts, distributed field...
Heterogeneous graph learning aims to capture complex relationships and diverse relational semantics among entities in a heterogeneous obtain meaningful representations for nodes edges. Recent advancements neural networks (HGNNs) have achieved state-of-the-art performance by considering relation heterogeneity using specialized message functions aggregation rules. However, existing frameworks limitations generalizing across datasets. Most of these follow the "pre-train" "fine-tune" paradigm on...
As a matter of fact, it is necessary for the manager to have better understanding supermarket sales nowadays. With this in mind, significant model analyze supermarkets help manager. On basis, article, XGBClassifier used do prediction sales. There are 750 training sets fit and 250 testing judge performance. The accuracy 0.49 train score about 62.67. normal way prediction, good data with big sets. This research helpful people who want manage supermarket, can give some important advice improve...
Digital agents for automating tasks across different platforms by directly manipulating the GUIs are increasingly important. For these agents, grounding from language instructions to target elements remains a significant challenge due reliance on HTML or AXTree inputs. In this paper, we introduce Aria-UI, large multimodal model specifically designed GUI grounding. Aria-UI adopts pure-vision approach, eschewing auxiliary To adapt heterogeneous planning instructions, propose scalable data...
Real-world data is represented in both structured (e.g., graph connections) and unstructured textual, visual information) formats, encompassing complex relationships that include explicit links (such as social connections user behaviors) implicit interdependencies among semantic entities, often illustrated through knowledge graphs. In this work, we propose GraphAgent, an automated agent pipeline addresses dependencies graph-enhanced inter-dependencies, aligning with practical scenarios for...
Modern recommender systems aim to deeply understand users' complex preferences through their past interactions. While deep collaborative filtering approaches using Graph Neural Networks (GNNs) excel at capturing user-item relationships, effectiveness is limited when handling sparse data or zero-shot scenarios, primarily due constraints in ID-based embedding functions. To address these challenges, we propose a model-agnostic recommendation instruction-tuning paradigm that seamlessly...
Smartphone agents are increasingly important for helping users control devices efficiently, with (Multimodal) Large Language Model (MLLM)-based approaches emerging as key contenders. Fairly comparing these is essential but challenging, requiring a varied task scope, the integration of different implementations, and generalisable evaluation pipeline to assess their strengths weaknesses. In this paper, we present SPA-Bench, comprehensive SmartPhone Agent Benchmark designed evaluate...
Knowledge Graphs (KGs) have emerged as invaluable resources for enriching recommendation systems by providing a wealth of factual information and capturing semantic relationships among items. Leveraging KGs can significantly enhance performance. However, not all relations within KG are equally relevant or beneficial the target task. In fact, certain item-entity connections may introduce noise lack informative value, thus potentially misleading our understanding user preferences. To bridge...
Self-supervised learning (SSL) has gained significant interest in recent years as a solution to address the challenges posed by sparse and noisy data recommender systems. Despite growing number of SSL algorithms designed provide state-of-the-art performance various recommendation scenarios (e.g., graph collaborative filtering, sequential recommendation, social KG-enhanced recommendation), there is still lack unified frameworks that integrate across different domains. Such framework could...
GNN-based recommenders have excelled in modeling intricate user-item interactions through multi-hop message passing. However, existing methods often overlook the dynamic nature of evolving interactions, which impedes adaption to changing user preferences and distribution shifts newly arriving data. Thus, their scalability performances real-world environments are limited. In this study, we propose GraphPro, a framework that incorporates parameter-efficient graph pre-training with prompt...
UAV (Unmanned Aerial Vehicle) swarm search has the advantages of flexible deployment, no casualties, and high cost-effectiveness. It become a force that cannot be ignored in battlefield. Aiming at task planning problem search, this paper treats each as subsystem based on self-organization idea, proposes algorithm IAPF (Improved Artificial Potential Field ). First, order to improve efficiency reduce computational complexity, new type target attraction field function was constructed....