Search

CTR prediction is essential for modern recommender systems. Ranging from early factorization machines to deep learning based models in recent years, existing CTR methods focus on capturing useful feature interactions or mining important behavior patterns. Despite the effectiveness, we argue that these methods suffer from the risk of label sparsity (i.e., the user-item interactions are highly sparse with respect to the feature space), label noise (i.e., the collected user-item interactions are usually noisy), and the underuse of domain knowledge (i.e., the pairwise correlations between samples). To address these challenging problems, we propose a novel Multi-Interest Self-Supervised learning (MISS) framework which enhances the feature embeddings with interest-level self-supervision signals. With the help of two novel CNN-based multi-interest extractors,self-supervision signals are discovered with full considerations of different interest representations (point-wise and union-wise), interest dependencies (short-range and long-range), and interest correlations (inter-item and intra-item). Based on that, contrastive learning losses are further applied to the augmented views of interest representations, which effectively improves the feature representation learning. Furthermore, our proposed MISS framework can be used as an plug-in component with existing CTR prediction models and further boost their performances. Extensive experiments on three large-scale datasets show that MISS significantly outperforms the state-of-the-art models, by up to 13.55% in AUC, and also enjoys good compatibility with representative deep CTR models., Accepted by ICDE2022

In traditional biped robot motion planning and control process, linear invert pendulum model is used to generate walking trajectory, with no consideration of the change of angular momentum, which sometimes cannot be ignored. In this paper, we propose a trajectory compensate method that takes into account the precise change of centroidal angular momentum during robot moving, including upper limb movement, foot swing and torso orientation movement, etc. We also provide a QP based formulation to solve the center of mass (CoM) trajectory from the Centroidal Moment Pivot (CMP) trajectory. We can adjust different properties of the CoM trajectory by adding different cost functions. The QP based formulation also have a bigger tolerance of trajectory error, which enhance the robustness of the solved CoM trajectory. Furthermore, we established an angular momentum preview control scheme, extend the method to not only offline trajectory planning, but also real-time trajectory control circumstances. We test our compensate method and control scheme in simulation and get obvious improvement in robot walking control.

Bipedal humanoid robots are unstable to external perturbations, especially when they are walking on uneven terrain in the presence of unforeseen collisions or obtained a external push force. The motion state capture point balance algorithm, need to adjust many hyper-parameters, and the parameters needed by the robot in different environments are nonlinear. We present a push recovery controller combined with reinforcement learning methond and get optimal parameters of the recovery controller. The experimental results show that our method can achieve the balance control of biped robot, and achieve good adaptability in simulation environment.