Content of A.pngicial Intelligence in our journal

Published in last 1 year |  In last 2 years |  In last 3 years |  All Please wait a minute... For Selected: Download Citations EndNote Ris BibTeX Toggle Thumbnails Select Groupwise Learning to Rank Algorithm with Introduction of Activated Weighting LI Yuxuan, HONG Xuehai, WANG Yang, TANG Zhengzheng, BAN Yan Journal of Frontiers of Computer Science and Technology    2022, 16 (7): 1594-1602.   DOI: 10.3778/j.issn.1673-9418.2101045 Learning to rank (LtR) applies supervised machine learning (SML) technologies to the ranking problems, aiming at optimizing the relevance of input document list. As regard to previous studies on the deep ranking model, the calculation of the relevance of the documents in the list is independent of each other, which lacks consideration of document interactions. In recent years, some new methods are devoted to mining the interaction between documents, such as groupwise scoring function (GSF), which learns multivariate scoring function to jointly judge the correlation, but most of these methods ignore the differences of the interaction between documents, and bring high calculation cost at the same time. In order to solve this problem, this paper proposes a weighted groupwise deep ranking model (W-GSF). In view of the deep interest network in the field of recommendation, this paper intro-duces the idea of adjusting the weight of historical behavior sequence according to the candidate products. On the basis of multivariate scoring method in learning to rank field, this method uses muti-layer feed forword neural networks as main structure, and adds an activation unit into it before the input module, taking advantage of neural networks to adjust the weight of input multiple variables adaptively, so as to mine the differences of cross document relationship. Experiments on the public benchmark dataset MSLR verify the effectiveness of the method. Compared with baseline ranking models, the introduction of activation strategy brings a significant improvement of ranking metrics, and the computational complexity is greatly reduced compared with the same effect learning to rank methods. Table and Figures | Reference | Related Articles | Metrics Abstract （267） PDF （198） HTML （8） Select Long Text Generation Adversarial Network Model with Self-Attention Mechanism XIA Hongbin, XIAO Yifei, LIU Yuan Journal of Frontiers of Computer Science and Technology    2022, 16 (7): 1603-1610.   DOI: 10.3778/j.issn.1673-9418.2104038 In recent years, the communication between human and computer has reached the inseparable degree, so the natural language processing as interaction technology between human and machine attracts more and more attention from researchers. Text generation is one of the common tasks of natural language processing. Currently, generative adversarial networks (GAN) is widely used in the field of text generation and the performance is excellent. To solve the sparse problem of scalar guided signals in traditional generative adversarial network discriminator and the limitation of learning only partial semantic information of text, a model based on the combination of multi-head self-attention leak generative adversarial networks (SALGAN) is proposed. Firstly, the feature vector is extracted by using the CNN model integrated with the multi-head self-attention mechanism as the feature extractor to enhance the feature extraction ability. Secondly, the features extracted by the discriminator are sent to the generator as step-by-step guidance signals to guide the generator to generate text, which makes the generated text more inclined to the reference text. Finally, the generator generates the text and passes it to the discriminator to determine whether it is true or not, in order to confirm whether the text meets the standards of human language. Experiments are carried out on two real datasets, COCO image captions and EMNLP2017 news, and the BLEU index is used for evaluation. The experimental results show that the text contains global semantic information after the multi-head self-attention mechanism is integrated into the CNN model, and the feature extraction performance of the CNN model is significantly improved. Table and Figures | Reference | Related Articles | Metrics Abstract （417） PDF （275） HTML （12） Select Dynamic Pickup-Point Recommendation Based on Spatiotemporal Trajectory and Hybrid Gain Evaluation GUO Yuhan, LIU Qiuyue Journal of Frontiers of Computer Science and Technology    2022, 16 (7): 1611-1622.   DOI: 10.3778/j.issn.1673-9418.2012039 With the rapid development of Internet technology, online car-hailing has become an important way of travel. Recommending boarding points for users through intelligent means can not only effectively achieve traffic diversion and alleviate congestion, but also reduce the communication cost between passengers and drivers, improve the service efficiency of drivers and reduce the waiting time of passengers, so as to improve the travel experience of both drivers and passengers. However, the existing recommendation methods are normally based on a single criterion, which do not achieve a good balance between passenger convenience and driver income, and can not guarantee the safety and accessibility of the recommended boarding point. By summarizing and analyzing the big data of spatiotemporal trajectory, this paper extracts the potential boarding points to ensure accessibility, avoids the bias caused by considering a single criterion, and comprehensively considers the key factors such as passenger walking income, driver driving income, road condition, and peripheral safety. It establishes a composite income evaluation of boarding points, and constructs the dynamic recommendation model of boarding points to control the recommended quantity of the same boarding point at the same time with constraints, effectively solving the unnecessary waiting and waste of resources caused by the accumulation of orders at a single boarding point, and alleviating the traffic pressure to a certain extent. Experiments based on real online car-hailing data show that the proposed model and recommendation method can achieve an effective dynamic allocation of boarding points, and have better comprehensive benefits and time advantages than the single criterion method. From the perspective of both drivers and passengers, on the basis of reducing the total travel time, the proposed method improves the overall driver pick-up efficiency and reduces the waiting time of passengers, and the comprehensive evaluation shows the obtained results are better than those provided by the existing recommendation methods. Table and Figures | Reference | Related Articles | Metrics Abstract （436） PDF （309） HTML （6） Select Hybrid Optimization Algorithm for Vehicle Routing Problem with Simultaneous Delivery-Pickup LI Jun, DUAN Yurong, HAO Liyan, ZHANG Weiwei Journal of Frontiers of Computer Science and Technology    2022, 16 (7): 1623-1632.   DOI: 10.3778/j.issn.1673-9418.2105024 In order to provide reasonable and effective decision support for logistics enterprises in vehicle distribu-tion path planning, this paper studies the vehicle routing problem with simultaneous delivery-pickup and time windows (VRPSDPTW) for single distribution center distribution mode, and establishes a mathematical model with the objective of minimizing the total distribution cost. According to the characteristics of the model, a hybrid optimization algorithm (SA-ALNS) based on the combination of simulated annealing (SA) and adaptive large-scale neighborhood search (ALNS) is proposed. An insertion heuristic algorithm based on time and distance weighting is used to construct the initial solution of the problem. A variety of delete and insert operators are introduced to optimize the path with adaptive selection strategy. Through the feedback mechanism, the probability of each operator being selected is gradually adjusted to make the algorithm more inclined to choose the operator with better optimization effect. The Metropolis criterion of simulated annealing mechanism is used to control the solution updating. In the simulation experiment, 56 large-scale examples are tested, and other intelligent optimization algori-thms such as p-SA algorithm, DCS algorithm and VNS-BSTS are compared and statistically analyzed. The results show that the algorithm is feasible and superior in solving the vehicle routing problem with simultaneous delivery-pickup and time windows. The research results greatly enrich the related research of vehicle routing problem (VRP). Table and Figures | Reference | Related Articles | Metrics Abstract （367） PDF （180） HTML （11）