A5090734254- Recommender Systems and Techniques
- Topic Modeling
- Domain Adaptation and Few-Shot Learning
- AI in cancer detection
- Advanced Graph Neural Networks
- Music and Audio Processing
- Landslides and related hazards
- Financial Distress and Bankruptcy Prediction
- Imbalanced Data Classification Techniques
- COVID-19 diagnosis using AI
- Time Series Analysis and Forecasting
- FinTech, Crowdfunding, Digital Finance
- Tree Root and Stability Studies
- Colorectal Cancer Screening and Detection
- Pancreatic and Hepatic Oncology Research
- Generative Adversarial Networks and Image Synthesis
- Power Quality and Harmonics
- Anomaly Detection Techniques and Applications
- Caching and Content Delivery
- Machine Learning in Healthcare
- Stock Market Forecasting Methods
- Power Transformer Diagnostics and Insulation
- Currency Recognition and Detection
- Machine Learning in Materials Science
- Image Processing and 3D Reconstruction
Central South University
2023-2024
The cross-domain recommendation (CDR) model addresses challenges such as data sparsity, the long tail distribution of user-item interactions, and cold start items or users. However, solely transferring domain-shared knowledge based on co-occurrence patterns, without considering user preferences, leads to negative transfer in CDR. To overcome these limitations, we propose an advanced deep learning CDR called Domain Adversarial Deep Interest Network (DADIN) aims facilitate smooth from source...
Cross-domain recommendation aims to leverage heterogeneous information transfers knowledge from a data-sufficient domain (source domain) data-scarce (target domain). Existing approaches mainly focus on learning single-domain user preferences and then employ transferring module obtain cross-domain preferences, but ignore the modeling of users' specific items. We argue that incorporating domain-specific source will introduce irrelevant fails target domain. Additionally, directly combining...
Click-Through Rate (CTR) prediction is one of the main tasks recommendation system, which conducted by a user for different items to give results. Cross-domain CTR models have been proposed overcome problems data sparsity, long tail distribution user-item interactions, and cold start or users. In order make knowledge transfer from source domain target more smoothly, an innovative deep learning cross-domain model, Domain Adversarial Deep Interest Network (DADIN) convert task into adaptation...
Establishing a reliable credit card fraud detection model has become primary focus for academia and the financial industry. The existing anti-fraud methods face challenges related to low recall rates, inaccurate results, insufficient causal modelling ability. This paper proposes based on counterfactual data enhancement of triplet network. Firstly, we convert problem generating optimal explanations (CFs) into policy optimization agents in discrete-continuous mixed action space, thereby...
Recommendation system algorithm based on multi-task learning (MTL) is the major method for Internet operators to understand users and predict their behaviors in multi-behavior scenario of platform. Task correlation an important consideration MTL goals, traditional models use shared-bottom gating experts realize shared representation information differentiation. However, The relationship between real-world tasks often more complex than existing methods do not handle properly sharing...
There are two fundamental problems in applying deep learning/machine learning methods to disease classification tasks, one is the insufficient number and poor quality of training samples; another how effectively fuse multiple source features thus train robust models. To address these problems, inspired by process human knowledge, we propose Feature-aware Fusion Correlation Neural Network (FaFCNN), which introduces a feature-aware interaction module feature alignment based on domain...
Landslide is a natural disaster that can easily threaten local ecology, people's lives and property. In this paper, we conduct modelling research on real unidirectional surface displacement data of recent landslides in the area propose time series prediction framework named VMD-SegSigmoid-XGBoost-ClusterLSTM (VSXC-LSTM) based variational mode decomposition, which predict landslide more accurately. The model performs well test set. Except for random item subsequence hard to fit, root mean...
Recommendation system algorithm based on multi-task learning (MTL) is the major method for Internet operators to understand users and predict their behaviors in multi-behavior scenario of platform. Task correlation an important consideration MTL goals, traditional models use shared-bottom gating experts realize shared representation information differentiation. However, The relationship between real-world tasks often more complex than existing methods do not handle properly sharing...
In this paper, we propose a framework for endoscopic image classification based on deep transfer learning (TL), specifically designed to address the unique challenges of medical classification. Our approach focuses performance convolutional neural network (CNN) model ResNet architecture. To further improve model's prediction accuracy and robustness, introduce two variants: ResNe-SE ResNet-CBAM, which incorporate Squeeze-Excitation Module Convolutional Block Attention Module, respectively....
There are two fundamental problems in applying deep learning/machine learning methods to disease classification tasks, one is the insufficient number and poor quality of training samples; another how effectively fuse multiple source features thus train robust models. To address these problems, inspired by process human knowledge, we propose Feature-aware Fusion Correlation Neural Network (FaFCNN), which introduces a feature-aware interaction module feature alignment based on domain...