Explore the vast range of titles available.

The driving behavior of bus drivers is related to the safety of all passengers and regulation of urban traffic. In order to analyze the relevant characteristics of speed and acceleration, accurate bus trajectories and patterns are essential for driver behavior analysis and development of effective intelligent public transportation. Exploiting real-time vehicle tracking, this paper develops a platform with vehicle-mounted terminals using differential global navigation satellite system (DGNSS) modules for driver behavior analysis. The DGNSS traces were used to derive the vehicle trajectories, which were then linked to road information to produce speed and acceleration matrices. Comprehensive field tests were undertaken on multiple bus routes in urban environments. The spatiotemporal results indicate that the platform can automatically and accurately extract the driving behavior characteristics. Furthermore, the platform’s visual function can be used to effectively monitor driving risks, such as speeding and fierce acceleration, in multiple bus routes. The details of the platform’s features are provided for intelligent transport system (ITS) design and applications.

The gelatinization and retrogradation characteristics of wheat starch affect the eating quality of Chinese-style food. Rapid Visco Analyzer (RVA) parameters have been widely used as important indicators to evaluate and improve the quality of wheat starch. However, the genetic basis of RVA parameters remains to be further explored. In the present study, a natural population was genotyped using 90K single nucleotide polymorphism (SNP) arrays, and the RVA parameters of this population grown in five environments were evaluated. The results showed that 22,068 high-quality SNP markers were identified and distributed unequally on the chromosomes. According to the genetic distance, 214 wheat materials were divided into four groups. Except for the pasting temperature (PTT), six parameters followed a normal distribution. Based on the general linear model, 969 significant association SNPs were detected by genome-wide association studies (GWAS), and chromosomes 7A and 2B had the most associated SNPs. Breakdown viscosity (BV) was associated with the most SNPs ( n = 238), followed by PTT ( n = 186), peak viscosity (PV; n = 156), trough viscosity (TV; n = 127), and final viscosity (FV; n = 126). According to the average linkage disequilibrium (LD), 33 stable quantitative trait loci (QTLs) were identified for single parameters in multiple environments, of which 12 were associated with BV, followed by peak time (PT; n = 8) and PTT ( n = 7). On the other hand, 67 pleiotropic QTLs were identified for multiple parameters. Three candidate genes— TasbeIIa, TasbeI , and TassIIa —were screened for phenotyping analysis. The grain width and the weight of the TasbeIIa and Ta SS IIa knockout (KO) lines were significantly lower than those of the TasbeI KO lines and the control (CK). The KO lines had smaller endosperm cells, smaller A-type starch granules, and higher amylose content. The TasbeI KO lines showed normal RVA curves, while the TasbeIIa KO lines showed flat curves. However, the TaSSIIa lines failed to paste under the RVA temperatures. Conclusively, the SNPs/QTLs significantly associated with the RVA parameters and genetic resources with novel haplotypes could be used to improve the quality of wheat starch.

Normally entrepreneur would raise fund from angel investors during the initial round. If the venture program was by then successful, the entrepreneur would then continue the fund-raising process from venture capitalist. By adopting the convertible preferred stock, we managed to construct the two-stage angel investment decision process. This research reveals the following: (1) The probability of the first stage’s success has negative relationships with levels of priority dividend in both first and second stages, as well as with the venture capitalist’s proportion of shares. (2) The probability of the second stage’s success has negative relationships with the venture capitalist’s proportion of shares and the dividend level of both first and second stage funding. (3) There has been a threshold of dividend distribution, which belongs to angel investor. While the level of angel investor’s shares is higher than the threshold, AN would decide to join the second phase of the program; otherwise, AN would exit the project at the end of the first stage.

This paper proposes a distributed containment constraint control algorithm considering the interference of ocean waves and actuator saturation. In this paper, an improved barrier Lyapunov function is introduced at the motion layer to construct a constraint boundary, thereby constraining the containment position errors. Additionally, a fixed-time interference observer is utilized to estimate the wave interferences. Meanwhile, a virtual control law is designed to eliminate containment position errors in real-time, enhancing the anti-interference ability and navigation reliability of the multi-unmanned surface vehicles (USVs) system. Based on the proof of the convergence of the system errors, the effectiveness of the control strategy is verified through numerical simulation experiments using the simuNPS software.

Modern listwise recommendation systems need to consider both long-term user perceptions and short-term interest shifts. Reinforcement learning can be applied on recommendation to study such a problem but is also subject to large search space, sparse user feedback and long interactive latency. Motivated by recent progress in hierarchical reinforcement learning, we propose a novel framework called mccHRL to provide different levels of temporal abstraction on listwise recommendation. Within the hierarchical framework, the high-level agent studies the evolution of user perception, while the low-level agent produces the item selection policy by modeling the process as a sequential decision-making problem. We argue that such framework has a well-defined decomposition of the outra-session context and the intra-session context, which are encoded by the high-level and low-level agents, respectively. To verify this argument, we implement both a simulator-based environment and an industrial dataset-based experiment. Results observe significant performance improvement by our method, compared with several well-known baselines. Data and codes have been made public.

Recommendation system has been a widely studied task both in academia and industry. Previous works mainly focus on homogeneous recommendation and little progress has been made for heterogeneous recommender systems. However, heterogeneous recommendations, e.g., recommending different types of items including products, videos, celebrity shopping notes, among many others, are dominant nowadays. State-of-the-art methods are incapable of leveraging attributes from different types of items and thus suffer from data sparsity problems. And it is indeed quite challenging to represent items with different feature spaces jointly. To tackle this problem, we propose a kernel-based neural network, namely deep unified representation (or DURation) for heterogeneous recommendation, to jointly model unified representations of heterogeneous items while preserving their original feature space topology structures. Theoretically, we prove the representation ability of the proposed model. Besides, we conduct extensive experiments on real-world datasets. Experimental results demonstrate that with the unified representation, our model achieves remarkable improvement (e.g., 4.1% ~ 34.9% lift by AUC score and 3.7% lift by online CTR) over existing state-of-the-art models.

Self-attention has become increasingly popular in a variety of sequence modeling tasks from natural language processing to recommendation, due to its effectiveness. However, self-attention suffers from quadratic computational and memory complexities, prohibiting its applications on long sequences. Existing approaches that address this issue mainly rely on a sparse attention context, either using a local window, or a permuted bucket obtained by locality-sensitive hashing (LSH) or sorting, while crucial information may be lost. Inspired by the idea of vector quantization that uses cluster centroids to approximate items, we propose LISA (LInear-time Self Attention), which enjoys both the effectiveness of vanilla self-attention and the efficiency of sparse attention. LISA scales linearly with the sequence length, while enabling full contextual attention via computing differentiable histograms of codeword distributions. Meanwhile, unlike some efficient attention methods, our method poses no restriction on casual masking or sequence length. We evaluate our method on four real-world datasets for sequential recommendation. The results show that LISA outperforms the state-of-the-art efficient attention methods in both performance and speed; and it is up to 57x faster and 78x more memory efficient than vanilla self-attention.

Feed recommendation allows users to constantly browse items until feel uninterested and leave the session, which differs from traditional recommendation scenarios. Within a session, user's decision to continue browsing or not substantially affects occurrences of later clicks. However, such type of exposure bias is generally ignored or not explicitly modeled in most feed recommendation studies. In this paper, we model this effect as part of intra-session context, and propose a novel intra-session Context-aware Feed Recommendation (INSCAFER) framework to maximize the total views and total clicks simultaneously. User click and browsing decisions are jointly learned by a multi-task setting, and the intra-session context is encoded by the session-wise exposed item sequence. We deploy our model online with all key business benchmarks improved. Our method sheds some lights on feed recommendation studies which aim to optimize session-level click and view metrics.

Sequential recommendation plays an increasingly important role in many e-commerce services such as display advertisement and online shopping. With the rapid development of these services in the last two decades, users have accumulated a massive amount of behavior data. Richer sequential behavior data has been proven to be of great value for sequential recommendation. However, traditional sequential models fail to handle users' lifelong sequences, as their linear computational and storage cost prohibits them from performing online inference. Recently, lifelong sequential modeling methods that borrow the idea of memory networks from NLP are proposed to address this issue. However, the RNN-based memory networks built upon intrinsically suffer from the inability to capture long-term dependencies and may instead be overwhelmed by the noise on extremely long behavior sequences. In addition, as the user's behavior sequence gets longer, more interests would be demonstrated in it. It is therefore crucial to model and capture the diverse interests of users. In order to tackle these issues, we propose a novel lifelong incremental multi-interest self attention based sequential recommendation model, namely LimaRec. Our proposed method benefits from the carefully designed self-attention to identify relevant information from users' behavior sequences with different interests. It is still able to incrementally update users' representations for online inference, similarly to memory network based approaches. We extensively evaluate our method on four real-world datasets and demonstrate its superior performances compared to the state-of-the-art baselines.

Graph neural networks (GNN) have been successful in many fields, and derived various researches and applications in real industries. However, in some privacy sensitive scenarios (like finance, healthcare), training a GNN model centrally faces challenges due to the distributed data silos. Federated learning (FL) is a an emerging technique that can collaboratively train a shared model while keeping the data decentralized, which is a rational solution for distributed GNN training. We term it as federated graph learning (FGL). Although FGL has received increasing attention recently, the definition and challenges of FGL is still up in the air. In this position paper, we present a categorization to clarify it. Considering how graph data are distributed among clients, we propose four types of FGL: inter-graph FL, intra-graph FL and graph-structured FL, where intra-graph is further divided into horizontal and vertical FGL. For each type of FGL, we make a detailed discussion about the formulation and applications, and propose some potential challenges.