Publications

Towards Informative and Diverse Dialogue Systems

Towards Informative and Diverse Dialogue Systems over Hierarchical Crowd Intelligence Knowledge Graph.

Predicts app popularity based on interactions and internal factors

DeePOP predicts app popularity based on interactions and internal factors, outperforming state-of-the-art methods.

Predicts app popularity based on interactions and internal factors

MetaDetector: Meta Event Knowledge Transfer for Fake News

MetaDetector uses adversarial adaptation to transfer meta knowledge between events for accurate detection of fake news. Outperforms state-of-the-art methods on real-world datasets.

MetaDetector: Meta Event Knowledge Transfer for Fake News

Learning Shared Mobility-aware Knowledge

MultiST is a framework that predicts multiple travel demands by leveraging shared and unique spatial-temporal knowledge using CNNs, GRUs, and a recurrent Gaussian cell.

Learning Dynamic Usage Graph for Mobile App Recommendation

DUGN models app usage dependencies and user interests for improved mobile app recommendations, outperforming existing methods.

Learning Dynamic Usage Graph for Mobile App Recommendation

Heterogeneous Sequence Modeling for Express Delivery Prediction

DeepExpress is a deep learning model that accurately predicts express delivery sequence by capturing feature-sequence couplings.

Heterogeneous Sequence Modeling for Express Delivery Prediction

Dynamic Probabilistic Graphical Model for Fake News Detection

It proposes a progressive fake news detection approach using LVDKF, which captures temporal information and utilizes crowd wisdom for better performance.

Dynamic Probabilistic Graphical Model for Fake News Detection

Determinants and effectiveness of misinformation correction

Investigation of the Determinants for Misinformation Correction Effectiveness on Social Media during COVID-19 pandemic.

Determinants and effectiveness of misinformation correction

CrowdOS:A Ubiquitous Operating System

CrowdOS:A Ubiquitous Operating System for Crowdsourcing and Mobile Crowd Sensing.

CrowdOS:A Ubiquitous Operating System

CrowdIM: Crowd-Inspired Intelligent Manufacturing Space Design

Crowd-inspired intelligent manufacturing space (CrowdIM)explores the use of natural crowd intelligence to enhance CrowdIM, proposing a general model and discussing it through a case study.

CrowdIM: Crowd-Inspired Intelligent Manufacturing Space Design

CrowdHMT: Crowd Intelligence with the Deep Fusion

CrowdHMT envisions a self-adaptive smart space with the deep fusion of human, machine, and IoT intelligence.

CrowdHMT: Crowd Intelligence with the Deep Fusion

Cross-FCL:Cross-edge Federated Continual Learning Framework

Cross-FCL:Toward a Cross-edge Federated Continual Learning Framework in Mobile Edge Computing Systems

Cross-FCL:Cross-edge Federated Continual Learning Framework

CausalSE: Varied Spatial Effects with Missing Data

CausalSE:Understanding Varied Spatial Effects with Missing Data Towards Adding New Bike-sharing Stations

CausalSE: Varied Spatial Effects with Missing Data

CAQ:Context-aware and Self-adaptive Deep Model Computation

CAQ:Towards Context-aware and Self-adaptive Deep Model Computation for AIoT Applications

CAQ:Context-aware and Self-adaptive Deep Model Computation

AppLife framework improves app survival prediction using MTL

Which App Is Going to Die? A Framework for App Survival Prediction with Multi-Task Learning".

Transfer Learning-Based Method for Surface Defect Detection

CTL-SDD:A Transfer Learning-Based Method for Surface Defect Detection with Few Samples.

Transfer Learning-Based Method for Surface Defect Detection

Task-Adaptative Meta-Learning Model for Store Placement

MetaStore:A Task-Adaptative Meta-Learning Model for Optimal Store Placement with Multi-City Knowledge Transfer.

Task-Adaptative Meta-Learning Model for Store Placement

Knowledge Transfer with Weighted Adversarial Network

Knowledge Transfer with Weighted Adversarial Network for Cold-Start Store Site Recommendation.

Knowledge Transfer with Weighted Adversarial Network

Joint Individual-level and Community-level Behavior Interaction

AskMe：uses rich hybrid behavior interactions to improve question recommendation performance in CQA sites.

Joint Individual-level and Community-level Behavior Interaction

Information-rich, logical dialogue systems

Knowledge-enhanced dialogue systems integrate external knowledge to improve response quality, but face challenges selecting, understanding, and integrating knowledge.

Information-rich, logical dialogue systems

Information-rich and Personalized Product Description Generation

CrowdDesigner generates superior personalized product descriptions for diverse applications.

DeepStore:An Interaction-aware Wide&Deep Model

DeepStore:An Interaction-aware Wide&Deep Model for Store Site Recommendation with Attentional Spatial Embeddings.

DeepStore:An Interaction-aware Wide&Deep Model

DeepDepict:Description Generation with the Deep Network

DeepDepict:Enabling Information Rich, Personalized Product Description Generation with the Deep Multiple Pointer Generator Network.

Crowdsensing 2.0

MCS is a user-driven sensing paradigm using smartphone sensors, enabling versatile applications like traffic planning and environment monitoring.

Crowdsensing 2.0

Cross-Domain Recommendation with Cross-Graph Knowledge Transfer Network

Cross-Domain Recommendation with Cross-Graph Knowledge Transfer Network.

Cross-Domain Recommendation with Cross-Graph Knowledge Transfer Network

Conditional Text Generation for Human-Machine Interaction

Knowledge Transfer with Weighted Adversarial Network for Cold-Start Store Site Recommendation.

Conditional Text Generation for Human-Machine Interaction

Spatial Community-Informed Evolving Graphs

Spatial Community-Informed Evolving Graphs for Demand Prediction", The European Conference on Machine Learning and Principles and Practice of Knowledge Discovery.

Spatial Community-Informed Evolving Graphs

Implicit human intelligence to understand crowdsourced data

From Crowdsourcing to Crowdmining:Using Implicit Human Intelligence for Better Understanding of Crowdsourced Data.

Implicit human intelligence to understand crowdsourced data

Decentralized AGV task allocation in multi-agent reinforcement learning

Decentralized Multi-AGV Task Allocation based on Multi-Agent Reinforcement Learning.

Decentralized AGV task allocation in multi-agent reinforcement learning

CrowdDepict:Generate Personalized Logical Product Description

CrowdDepict:Know What and How to Generate Personalized and Logical Product Description using Crowd Intelligence

CrowdDepict:Generate Personalized Logical Product Description

Citywide Fire Risk Forecasting Using A Machine Learning Approach

CityGuard:Citywide Fire Risk Forecasting Using A Machine Learning Approach.

Citywide Fire Risk Forecasting Using A Machine Learning Approach