|
★★★★★实验室学术交流平台:(https://www.neusncp.com)★★★★★
研究领域及方向:复杂网络/图神经网络、图强化学习、数字图像、工业大数据
研究方向1:复杂网络/图神经网络
1. 复杂网络理论及应用
面向复杂图结构数据的生成机制、演化过程、拓扑特征等理论研究及应用。
2. 多层网络信息挖掘理论及应用
面向多层网络的融合机制、节点影响力识别、社团检测等理论研究及应用
3. 图神经网络理论及应用
面向图神经网络的架构设计和优化、自监督学习、异构图学习等理论研究及应用
研究方向2:图强化学习
1. 基于深度强化学习的图结构数据分析
使用深度强化学习方法的复杂图结构数据建模、数据增强等理论研究及应用
2. 基于深度强化学习的图表示学习
使用深度强化学习方法的子图结构识别、链路预测、节点分类等理论研究及应用
3. 基于深度强化学习的知识图谱推理
使用强化学习方法的知识图谱建模、推理、推荐等理论研究及应用
研究方向3:工业大数据分析/电力AI
1. 面向工业生产企业的数字孪生技术
基于计算机图形学、人工智能等技术的面向工业企业全生命周期数字孪生技术研究及应用
2. 基于图像分割的工业生产图纸管理
基于图像分割的工业企业生产图纸分割、识别、分类等理论研究及应用
3. 基于时序预测的工业智能数据分析
基于时序预测的工业企业生产、销售数据的数据挖掘、数据预测等理论研究及应用
4. 基于深度学习/大模型的电力AI应用
基于电力大数据,运用深度学习和大模型技术助力AI在电力能源领域应用落地
研究方向4:数字图像
1. 数字图像安全存储技术
(1)面向数字图像的保密性研究:对典型的置乱-扩散架构的混沌图像加密算法的密码分析;利用S-box对基于DNA编码技术的图像加密算法的密码分析;
(2)面向数字图像的完整性研究:针对特定篡改方式的检测;通用的篡改检测。
2. 创意设计场景中交互式图像生成方法和解释性
聚焦创意设计场景中交互式图像生成的新需求,重点解决现有方法面临的可交互、可控、可解释性三个核心问题:
(1)基于扩散模型构建多任务兼容的图像生成框架;
(2)多条件可控的图像生成方法;
(3)扩散模型可解释性的文本表达方法。
Research Directions & Interests
1. Research on Complex Networks/Graph Neural Networks
(1) Theories and Applications of Complex Networks
(2) Theories and Applications of Multi-layer Network Information Mining
(3) Theories and Applications of Graph Neural Networks
2. Research on Graph Reinforcement Learning
(1) Theories and Applications of Graph-structured Data Analysis via Deep Reinforcement Learning
(2) Theories and Applications of Graph Representation Learning via Deep Reinforcement Learning
(3) Theories and Applications of Knowledge Graph Reasoning via Deep Reinforcement Learning
3. Industrial Big Data Analysis / Power AI
(1)Digital Twin Technology for Industrial Production Enterprises
Research and application of digital twin technology for the whole life cycle of industrial enterprises based on technologies such as computer graphics and artificial intelligence.
(2) Management of Industrial Production Drawings Based on Image Segmentation
Theoretical research and application on the segmentation, recognition and classification of industrial enterprise production drawings based on image segmentation.
(3) Industrial Intelligent Data Analysis Based on Time Series Prediction
Theoretical research and application on data mining and data prediction of industrial enterprise production and sales data based on time series prediction.
(4) Power AI Applications Based on Deep Learning / Large Models
Based on power big data, apply deep learning and large model technologies to help AI applications land in the power and energy field.
4. Digital Image
(1)Digital Image Security
l Image encryption and cryptanalysis: cryptanalysis of the permutation-diffusion image encryption; universal cryptanalysis of the DNA-based image encryption.
l Blind image forensics: image forgery detection for copy-move, removal and splicing; universal image forgery detection.
(2)Interactive image generation methods and interpretability in creative design scenarios
l Multi Task Compatible Image Generation Framework Based on Diffusion Model
l Theory and Application of Image Generation by Using Multi Condition Control
l Explainable Text Expression Methods of Diffusion Model
4. Research on Industrial Big Data
(1) Digital Twin Technology for Industrial Manufacturing Companies
(2) Management of Industrial Manufacturing Drawings Based on Image Segmentation
(3) Industrial Intelligence Data Analysis Based on Time Series Prediction

