Showing papers by "Xiaoou Li published in 2020"

Deformable Convolutional Networks for Efficient Mixed-Type Wafer Defect Pattern Recognition

Abstract: Defect pattern recognition (DPR) of wafer maps is critical for determining the root cause of production defects, which can provide insights for the yield improvement in wafer foundries. During wafer fabrication, several types of defects can be coupled together in a piece of wafer, it is called mixed-type defects DPR. To detect mixed-type defects is much more complicated because the combination of defects may vary a lot, from the type of defects, position, angle, number of defects, etc. Deep learning methods have been a good choice for complex pattern recognition problems. In this article, we propose a deformable convolutional network (DC-Net) for mixed-type DPR (MDPR) in which several types of defects are coupled together in a piece of wafer. A deformable convolutional unit is designed to selectively sample from mixed defects, then extract high-quality features from wafer maps. A multi-label output layer is improved with a one-hot encoding mechanism, which decomposes extract mixed features into each basic single defect. The experiment results indicate that the proposed DC-Net model outperforms conventional models and other deep learning models. Further results of the interpretable analysis reveal that the proposed DC-Net can accurately pinpoint the defects areas of wafer maps with noise points, which is beneficial for mixed-type DPR problems.

Redundant Robot Control Using Multi Agent Reinforcement Learning

Abstract: Robot control in task-space1 needs the inverse kinematics and Jacobian matrix. They are not available for redundant robots, because there are so many degrees-of-freedom (DOF). Intelligent learning methods, such as neural networks (NN) and reinforcement learning (RL) can learn them. However, NN needs big data and RL is not suitable for multilink robots as the redundant robots. In this paper, we propose a full cooperative multi-agent reinforcement learning (MARL) to solve the above problems. Each joint of the robot is regarded as one agent. Although the dimension of the learning space is very large, the full cooperative MARL uses the kinematic learning and avoids the function approximators in large learning space. The experimental results show that our MARL is much more better compared with the classic methods such as, Jacobian-based methods and neural networks.1Task-space (or Cartesian space) is defined by the position and orientation of the end effector of a robot. Joint-space is defined by angular displacements of each joint of a robot.

A Novel Fuzzy System With Adaptive Neurons for Earthquake Modeling

Abstract: Data driven fuzzy neural networks have some disadvantages, such as high dimensions and complex learning process. Also, the obtained models are difficult to interpret. In this paper, we propose a novel simple fuzzy system, which uses fuzzy adaptive neurons. This novel model takes the advantages of the interpretability of the fuzzy system and good approximation ability of the neural networks. We propose a simple learning algorithm for the novel fuzzy system. The stability analysis is given. We successfully apply this fuzzy model for the earthquake modeling. Comparisons with the popular fuzzy neural model are proposed.

Display Name-Based Anchor User Identification across Chinese Social Networks

Abstract: Anchor user identification across social networks is a classification task which determines whether a pair of accounts from different social networks belong to the same user. It is a fundamental research of information dissemination across social networks. Based on the observation that users prefer to use similar or identical display names in different social network, some researchers utilized the similarity between display names to build models. However, due to Chinese social network setting and pronunciation and font characteristics of Chinese display names, these methods do not perform well in Chinese social network datasets. To address this problem, we analyze the display name pairs of Chinese anchor users which are obtained by a crawler build in this paper. Then we define 4 special features to extract the pronunciation and font similarities. Finally, we use Gradient Boosting to establish the identification model. The experiments based on the ground-truth datasets we obtained show that these features can improve the performance of display name-based anchor user identification between Chinese social networks.

Multi-robot manipulation using formation control and human-in-the-loop scheme

Abstract: In this paper, we present a novel multi-robot manipulation system with human operator. We modify the formation control law, and apply it to the human-in-the-loop (HITL) scheme. We add a derivative term to the common formation control law to reduce the formation breakups in the transient state, which is one of difficult problems in the formation control. The HITL system is used to lead the formation while providing haptic feedback, such that the human can move the multi-robot with a desired prescribed formation. The method developed is applied to a platform to test the effectiveness and efficiency of the algorithm.

Robust Control in the Worst Case Using Continuous Time Reinforcement Learning

Abstract: Reinforcement learning (RL) is an effective method to design robust control. Uncertainty in the worst case requires large state-action learning space. The continuous time RL can solve this computational problem.In this paper, we modify the classical continuous time RL. Compared with the actor-critic (AC) algorithm, our method is more simple and more robust under the worst-case uncertainty.