Search

Training a dialogue policy using deep reinforcement learning requires a lot of exploration of the environment. The amount of wasted invalid exploration makes their learning inefficient. In this paper, we find and define an important reason for the invalid exploration: dead-ends. When a conversation enters a dead-end state, regardless of the actions taken afterward, it will continue in a dead-end trajectory until the agent reaches a termination state or maximum turn. We propose a dead-end resurrection (DDR) algorithm that detects the initial dead-end state in a timely and efficient manner and provides a rescue action to guide and correct the exploration direction. To prevent dialogue policies from repeatedly making the same mistake, DDR also performs dialogue data augmentation by adding relevant experiences containing dead-end states. We first validate the dead-end detection reliability and then demonstrate the effectiveness and generality of the method by reporting experimental results on several dialogue datasets from different domains.

Nowadays, social media plays an important role in many fields, such as the promotion of measures against major infectious diseases, merchandising, etc. In social media, some people are known as opinion leaders due to their strong ability to influence the opinions of others. The detection of opinion leaders has become an important task in social network analysis. Social networks are often represented in the form of graphs which allows a large number of graph analysis methods to be used for opinion leader detection. Some studies have attempted to apply graph representation learning for opinion leader detection and achieved good results. In this paper, we propose a model-agnostic framework that formulate the opinion leader detection problem as a ranking task of node embeddings. A variety of methods and datasets are chosen to analyze the performance of our framework both qualitatively and quantitatively. Based on the analysis results, we propose a strategy that combines opinion leaders detected by two different ranking algorithms to obtain a more comprehensive set of opinion leaders. And we analyze the temporal changes of the opinion leaders in one of the dynamic social networks.

Simultaneous Localization and Mapping (SLAM) is a fundamental problem for autonomous mobile robots (AMRs). AMRs are widely used in automated warehousing, factory material transfer systems, flexible assembly systems, and other intelligent transportation sites. The visual Inertial Odometry (VIO) which consists of the camera and inertial-measurement-unit (IMU), is a popular approach to achieve accurate 6-DOF state estimation. However, such locally accurate VIO is prone to drift and cannot provide a global consistent map. The prerequisite for re-localizing the robot and ensuring precise autonomous navigation is an accurate and global consistent map of its environment. In this study, we propose a stereo visual-inertial mapping system. The front-end is a robust stereo VIO based on a tightly-coupled sliding window optimization. The core of the back-end is the global Bundle-Adjustment (BA) which is a nonlinear optimization, in which IMU is also added as a time-domain constraint. Meanwhile, stereo-camera-IMU extrinsic calibration is performed in BA to improve mapping accuracy. The selection principles of keyframes and map points are also designed according to the AMRs application characteristics. Further, the forward and backward Perspective-n-Point (PNP) method is also adopted to avoid the loop-detection mismatch. The performance of the system was validated and compared against other state-of-the-art algorithms. The findings revealed the effectiveness and robustness of this stereo visual-inertial mapping algorithm.

Traditional Chinese medicine (TCM) syndromes of heart failure (HF) are considered as the key to realize the integration of disease and syndrome differentiation, as well as the correspondence between recipes and syndromes. The evolution of HF syndromes starts with changing of yin-yang and ends with collapsing of yin-yang. This process depends on five viscera, the center of which is the heart and kidneys. Then we present a hypothesis of yin-yang and five viscera on the evolution of HF syndromes. Standardization of HF syndromes is urgently needed, while methodology sheds light on the establishment and usage of standardization of HF syndromes in future. Objective of HF syndromes may lead to the discovery of the syndromes essence, which relies on case-based TCM evidence via integration of modern science and technology.

This work aimed to discuss the relationship between the Gensini score of blood-stasis syndrome (BSS) in coronary heart disease (CHD) and vascular endothelial growth factor (VEGF) levels. Ninety-six patients with BSS in CHD were divided into 48 cases with qi deficiency and blood-stasis pattern and 48 cases of intermingled phlegm and blood-stasis pattern, based on traditional Chinese medicine (TCM) syndrome differentiation. Coronary angiography was carried out, and VEGF level was determined. Simultaneously, the extent of blood stasis was assessed with the blood-stasis score. Furthermore, the result of coronary angiography was assessed with the Gensini score system. The result showed that the blood-stasis score and coronary artery stenosis in the group with qi deficiency and blood stasis were higher than those in the intermingled phlegm and blood-stasis group. The levels of serum VEGF and endothelin1 increased with the increase of the blood-stasis and Gensini scores, whereas the level of insulin-like growth factor-1 decreased with the increase of the blood-stasis and Gensini scores. Thus, the coronary Gensini score is in correlation with the VEGF level and blood-stasis score.