Recommender systems are increasingly prevalent to provide personalized suggestions and enhance user satisfaction. Typical recommendation models encode users and items as embeddings, and generate recommendations by assessing the similarity between these embeddings. Despite their effectiveness, these embedding-based models struggle with modeling user uncertainty and capturing diverse user interests using a single fixed user embedding. Recent studies have begun to explore a user-distribution paradigm to learn distributions for users. However, this approach employs a single distribution per user, which fails to effectively delineate semantic boundaries, resulting in sub-optimal recommendations. To this end, we propose GCDR, a Guided Conditional Diffusion Recommender model, to learn multiple distributions for each user in this paper. Specifically, GCDR addresses two major challenges: 1) learning disentangled distributions, and 2) learning personalized distributions. GCDR captures inter-user and intra-user distribution properties through conditional and guided diffusion, respectively. It maintains user-specific embeddings to encode long-term interests for conditional diffusion, while for guided diffusion, it incorporates short-term interests encoded from recent interactions with category preferences. To align the diffusion model with the recommendation task, we train GCDR with three loss functions, included the user loss, the recommendation loss and the diffusion loss. Extensive experiments on four real-world datasets show that GCDR is able to learn effective user distributions and is superior to thirteen state-of-the-art baseline methods.

Automotive energy management systems (EMSs) are evolving to achieve intelligence across their entire lifecycle, from initial product development to real-world customer usage. This paper proposes a bi-level transfer approach with model-agnostic meta-learning (MAML) to realize cross-platform transferable and online-adaptive EMS. In the development phase, MAML calibrates the heuristic control maps of an instantaneous optimization-based EMS, with a per-unit state-action space design ensuring seamless knowledge transfer between vehicle platforms. During the usage phase, real driving data are adopted to refine onboard control parameters. The framework is validated through real vehicle experiments. Firstly, the entire MAML-assisted V-cycle development is implemented to validate the optimality and knowledge transfer of the EMS, resulting in zero-shot transfer for EMS calibration on new vehicle products. Additionally, real vehicle experimental tests show that a correction of 8.0%~9.5% fuel economy is improved against the convention reinforcement learning-based EMS during usage via online-adaptation, effectively bridging the gap between pre-trained policies and real-world optimal energy management.

Recommender systems have emerged as an indispensable mean to meet personalized interests of users and alleviate information overload. Despite the great success, accuracy-oriented recommendation models are creating information cocoons, i.e., it is becoming increasingly difficult for users to see other items they might be interested in. Although recent studies start paying attention to enhancing recommendation diversity, models based on point embedding fail to describe the range of user preferences and item features well, which is essential for diversified matching. To this end, we propose LCD-UC , a novel List-Check-Decide framework with UnCertainty masking based on box embedding to improve recommendation diversity with recommendation accuracy maintained. Specifically, LCD-UC creates hypercubes to represent users and items using box embedding for high model flexibility and expressiveness. Then, a hypercube similarity scoring function is designed to measure the similarity between hypercubes representing users and items. To make a balance between the accuracy and diversity of recommendations and achieve personalized diversity needs, we further develop a user-item pairwise attention mechanism as well as a user uncertainty masking mechanism in LCD-UC. Besides, we present two new metrics for better evaluation on recommendation diversity, which address the issue that existing metrics only consider the coverage of categories while ignore the frequency of categories. The extensive experiments on three real-world datasets show that LCD-UC can improve both recommendation accuracy and diversity over three base models, and is superior to six state-of-the-art recommendation models. An online 10-day AB test also demonstrates that LCD-UC can improve the performance of a real-world advertising system.

Ridesharing service, as a sustainable transportation mode, has gained great interest in both industry and academic fields. Existing TNC ridesharing services provide prescriptive solutions without coordinating riders’ interactions and intentions on route choices. As a result, ridesharing services often end with ride-hailing services, and they are still under a relatively low usage rate. Motivated by this view, this study developed a real-time coordinated ridesharing route choice mechanism (CSM) for guiding riders’ ridesharing route choices to fill this research and application gap. Specifically, this study modeled this CSM as a pure-strategy atomic fare-sharing game based on the assumption that every rider is selfish and tries to choose the best route to minimize his/her travel fare among multiple feasible candidate routes. An existing tree-generation algorithm was used to find the candidate routes for each rider. This study proved the existence of a Nash Equilibrium in this game by constructing a potential function and proving this game is a potential game. This study further developed a sequential updated distributed algorithm and proved its convergence to explore an equilibrium solution of the CSM. To address the scalability issue, this study created a coalition formation approach based on ridesharing potential, to separate riders into ridesharing coalitions and then independently implement the CSM for each ridesharing coalition. Our experiments illustrate that the coalition approach scales down the problem size of each CSM and dramatically improves the computation efficiency while maintaining the same level of system performance. More importantly, the CSM can significantly promote riders’ usage of the ridesharing service by greatly saving their travel fares while satisfying their trip requirements.

Port traffic flow modeling based on big data is an important research direction in the shipping field, having the task of traffic forecasting for ports worldwide. Graph neural networks have a strong ability to capture the spatial topology characteristics and may be combined with recurrent neural networks or dilated convolution methods in time series prediction, producing a large number of spatiotemporal graph convolution models. Such models have been widely and successfully applied in traffic forecasting. Differing from urban traffic flow data, the statistical time span of port vessel flow and throughput data is large, its spatial span is wide, and the data experience significant fluctuations. Consequently, certain spatiotemporal graph convolution traffic prediction models are unsuitable for shipping scenarios. To address this shortcoming, we have created a unique port flow dataset based on automatic identification system (AIS) and port geographic data. Using theoretical analysis and experimental comparison, we have determined the most appropriate model for shipping predictions based on existing spatiotemporal graph models and have proposed model optimization recommendations for the maritime domain. Our experiment based on an open source traffic forecasting framework to compare the results of multiple existing spatiotemporal graph models under fair conditions with the central ports of Rotterdam, Shanghai, Boston, and Singapore. The results show that Graph WaveNet exhibits better performance in shipping scenarios.

Carpooling service obtains a significant interest in both industry and academic fields for its potential to improve mobility and save travel costs. However, the existing approaches for exploring online carpooling routes for riders struggle with the trade-off between system optimality and computation efficiency. This study is motivated to develop a semi-centralized ride-matching strategy (SCM) to address this difficulty. Specifically, we model the carpooling service problem (CSP) by a mixed-integer programming (CSP-MIP), which explores an optimal carpooling solution for a large-scale of riders to minimize system service time by using a small CAV fleet to serve all the riders. To address the computation difficulty, we first analyze and quantify the carpooling chances (C2) among riders according to their trip features. According to the C2, we decompose a large-scale CSP-MIP involving all riders in a network into small sub-MIPs, each including a small number of riders in a carpooling community, without over sacrificing the system optimality. For each sub-MIP, we develop a network flow algorithm combined with a greedy ride-matching model to explore a local optimal solution. The experiments built upon the Hardee network demonstrated that the SCM could solve the CSP efficiently. It significantly improves the computation efficiency while maintaining the system performance as it’s compared to the existing approaches.

Convolution neural network (CNN), which can extract rich semantic information of signal, is a representative feature learning network in acoustic scene classification (ASC). However, since that the receptive field (RF) of a CNN is fixed, it is inefficient to capture the dynamical time-frequency changing characteristic of the input Log-Mel spectrogram. In addition, although the Log-Mel spectrogram can be treated as an image, the time and frequency dimensions, which respectively represent the acoustic event duration and frequency information, have different physical meanings. Therefore, existing receptive field adaptive methods, which get same-sized optimal receptive fields in two dimensions, are not suitable for ASC. To tackle this problem, we propose a convolution receptive field dual selection mechanism (CRFDS) in this paper. In CRFDS, a multi-branch dual convolutional selection unit is designed to replace the convolution layers of CNN. The multi-branch dual convolutional selection unit is composed by a m-branch n × 1 convolution kernels concatenating with another m-branch 1 × n convolution kernels where n is a variable parameter with a value of 3, 5, …, 2 × m+1. A strong attention selection operat ion is embedded after each of the m-branch convolution kernels respectively to chose an optimal convolutional kernel size during learning. Because two strong attention selection operations in the unit are able to choose convolutional kernels with different sizes, the receptive field in two dimensions of our proposed CRFDS can be adjusted respectively. As a result, CRFDS has the ability to find the optimal RFs in two dimensions simultaneously, which improve the semantic feature extraction ability of CNN in spectrogram. We conduct acoustic scene classification experiments on the TUT Urban Acoustic Scenes 2019 development dataset, and the experimental results show that CRFDS can improve accuracy of ASC by 2.4%, which corroborate the performances of proposed mechanism.