Non-IID recommender system discloses the nature of recommendation and has shown its potential in improving recommendation quality and addressing issues such as sparsity and cold start. It leverages existing work that usually treats users/items as independent while ignoring the rich couplings within and between users and items, leading to limited performance improvement. In reality, users/items are related with various couplings existing within and between users and items, which may better explain how and why a user has personalized preference on an item. This work builds on non-IID learning to propose a neural user-item coupling learning for collaborative filtering, called CoupledCF. CoupledCF jointly learns explicit and implicit couplings within/between users and items w.r.t. user/item attributes and deep features for deep CF recommendation. Empirical results on two real-world large datasets show that CoupledCF significantly outperforms two latest neural recommenders: neural matrix factorization and Google's Wide&Deep network.

Making use of the expansion in a power series, the exact eigensolutions of two electrons in a parabolic quantum dot are obtained. The quantum-size effects on the energy spectra of two electrons are shown for the first time.

No abstract

In this paper, we first introduce an alternative proof of the error estimates of the numerical methods for solving linear fractional differential equations proposed in Diethelm [6] where a first-degree compound quadrature formula was used to approximate the Hadamard finite-part integral and the convergence order of the proposed numerical method is O(∆t 2−α ), 0 < α < 1, where α is the order of the fractional derivative and ∆t is the step size. We then use the similar idea to prove the error estimates of a high order numerical method for solving linear fractional differential equations proposed in Yan et al. [37], where a second-degree compound quadrature formula was used to approximate the Hadamard finite-part integral and we show that the convergence order of the numerical method is O(∆t 3−α ), 0 < α < 1. The numerical examples are given to show that the numerical results are consistent with the theoretical results.

Purpose This paper aims to resolve issues of the traditional artificial potential field method, such as falling into local minima, low success rate and lack of ability to sense the obstacle shapes in the planning process. Design/methodology/approach In this paper, an improved artificial potential field method is proposed, where the object can leave the local minima point, where the algorithm falls into, while it avoids the obstacle, following a shorter feasible path along the repulsive equipotential surface, which is locally optimized. The whole obstacle avoidance process is based on the improved artificial potential field method, applied during the mechanical arm path planning action, along the motion from the starting point to the target point. Findings Simulation results show that the algorithm in this paper can effectively perceive the obstacle shape in all the selected cases and can effectively shorten the distance of the planned path by 13%–41% with significantly higher planning efficiency compared with the improved artificial potential field method based on rapidly-exploring random tree. The experimental results show that the improved artificial potential field method can effectively plan a smooth collision-free path for the object, based on an algorithm with good environmental adaptability. Originality/value An improved artificial potential field method is proposed for optimized obstacle avoidance path planning of a mechanical arm in three-dimensional space. This new approach aims to resolve issues of the traditional artificial potential field method, such as falling into local minima, low success rate and lack of ability to sense the obstacle shapes in the planning process.