Lip Localization via Semi-Local Kernels – The paper presents a practical and robust method for learning and computing face models in the presence of natural occlusion. Our algorithm is based on a discriminative representation over faces, which is an essential step to learning the structure of a face database. We prove that both the face recognition and face estimation are NP-hard, without taking into account the presence of occlusion. We apply our method to several complex face datasets and show results on simulated and real-world datasets.
The success of deep reinforcement learning (RL) is largely due to the high computational cost of the RL algorithms. In this paper we compare the effectiveness of a well-known RL algorithm named Long Short-Term Memory (LSTM) with an expensive RL algorithm. We propose an efficient RL algorithm called Long Short-Term Memory RL (LSTM-RL), and show that LSTM-RL outperforms the current state-of-the-art RL methods for various tasks. We also show that it is a good value for evaluating RL algorithms in terms of the efficiency.
A Unified Approach for Online Video Quality Control using Deep Neural Network Technique
Directional Nonlinear Diffractive Imaging: A Review
Lip Localization via Semi-Local Kernels
Examining Kernel Programs Using Naive Bayes
Interactive Stochastic LearningThe success of deep reinforcement learning (RL) is largely due to the high computational cost of the RL algorithms. In this paper we compare the effectiveness of a well-known RL algorithm named Long Short-Term Memory (LSTM) with an expensive RL algorithm. We propose an efficient RL algorithm called Long Short-Term Memory RL (LSTM-RL), and show that LSTM-RL outperforms the current state-of-the-art RL methods for various tasks. We also show that it is a good value for evaluating RL algorithms in terms of the efficiency.