A5046681103- Robotics and Sensor-Based Localization
- Video Surveillance and Tracking Methods
- Data Stream Mining Techniques
- Infrared Target Detection Methodologies
- Advanced Graph Neural Networks
- Recommender Systems and Techniques
- Advanced Vision and Imaging
- Advanced Bandit Algorithms Research
- Expert finding and Q&A systems
- Caching and Content Delivery
- Data Mining Algorithms and Applications
- Customer churn and segmentation
- Machine Learning and Data Classification
Merck (Singapore)
2024
Alibaba Group (China)
2022
Harbin Institute of Technology
2017-2018
In recent years, recommender systems have advanced rapidly, where embedding learning for users and items plays a critical role. A standard method learns unique vector each user item. However, such has two important limitations in real-world applications: 1) it is hard to learn embeddings that generalize well with rare interactions; 2) may incur unbearably high memory costs when the number of scales up. Existing approaches either can only address one or flawed overall performances. this...
In recent years, recommender systems have advanced rapidly, where embedding learning for users and items plays a critical role. A standard method learns unique vector each user item. However, such has two important limitations in real-world applications: 1) it is hard to learn embeddings that generalize well with rare interactions; 2) may incur unbearably high memory costs when the number of scales up. Existing approaches either can only address one or flawed overall performances. this...
In this paper, a deep learning network is investigated to detect moving targets for UAV equipped with monocular camera. An algorithm based on fully convolutional proposed obtain the position and direction of targets. A Kalman filter incorporated into increase accuracy target information acquisition. The experimental results show effectiveness relatively low hardware resource consumption.
For years, machine learning has become the dominant approach to a variety of information retrieval tasks. The performance algorithms heavily depends on their hyperparameters. It is hence critical identity optimal hyperparameter configuration when applying algorithms. Most existing optimization methods assume static relationship between and algorithmic are thus not suitable for many applications with non-stationary environments such as e-commerce recommendation online advertising. To address...
In this paper, the information acquisition problem of columnar moving obstacles is investigated for a quadrotor equipped with stereo vision system flying in an indoor GPS-denied environment. A algorithm based on proposed to obtain position and velocity direction. The spatial relation between object detected used remove interferences which features are similar depth image. Kalman Filter incorporated into increase accuracy obstacle information. Experimental results show effectiveness low...
In this paper, a moving target detection problem is investigated for the unmanned air vehicles with monocular vision sensor. First, algorithm based on U-Net network proposed to obtain position and velocity direction of target. Then, Kalman Filter incorporated into increase accuracy states acquisition targets. experimental results show has low computation better performance.