A5103166791- Supply Chain and Inventory Management
- Multimodal Machine Learning Applications
- Advanced Queuing Theory Analysis
- Sustainable Supply Chain Management
- Advanced Optimization Algorithms Research
- Advanced Image and Video Retrieval Techniques
- Auction Theory and Applications
- Domain Adaptation and Few-Shot Learning
- Consumer Market Behavior and Pricing
- Advanced Neural Network Applications
- Face recognition and analysis
- Digital Platforms and Economics
- Optimization and Variational Analysis
- Robotics and Sensor-Based Localization
- Generative Adversarial Networks and Image Synthesis
- Face and Expression Recognition
- Optimization and Search Problems
- Scheduling and Optimization Algorithms
- Optimization and Packing Problems
- Supply Chain Resilience and Risk Management
- 3D Surveying and Cultural Heritage
- Vehicle Routing Optimization Methods
- Sharing Economy and Platforms
- Advanced Image Processing Techniques
- Fixed Point Theorems Analysis
Liaoning University of Technology
2025
First Affiliated Hospital Zhejiang University
2025
National Earthquake Response Support Service
2023-2025
University of Science and Technology of China
2021-2024
University of Chinese Academy of Sciences
2022-2024
Fudan University
2013-2024
Space Engineering University
2024
Huazhong University of Science and Technology
2007-2024
Shanghai Jiao Tong University
2022-2024
Institute of Automation
2021-2024
Traditionally, online retailers have acted as product resellers. Recently, these also started to serve marketplaces by providing a platform directly connect sellers with buyers. Over and above re‐shaping the traditional e‐commerce market, conventional wisdom suggests that this new format will mitigate double‐marginalization effect benefit both intermediary suppliers through revenue sharing scheme. However, we find upstream competition between critically moderates possibility. We interaction...
In this paper, we present Co-scale conv-attentional image Transformers (CoaT), a Transformer-based classifier equipped with co-scale and mechanisms. First, the mechanism maintains integrity of Transformers’ encoder branches at individual scales, while allowing representations learned different scales to effectively communicate each other; design series serial parallel blocks realize mechanism. Second, devise by realizing relative position embedding formulation in factorized attention module...
This paper studies information sharing in a distribution channel where the manufacturer possesses better demand-forecast than downstream retailer. We examine three information-sharing formats: no (i.e., ex ante commits to not its forecast), voluntary makes decision post after receiving and mandatory is mandated share forecast). characterize equilibrium outcomes under formats investigate firms’ preferences regarding these formats. It shown that when retailer risk-neutral, both firms are...
In this paper, we present a regression-based pose recognition method using cascade Transformers. One way to categorize the existing approaches in domain is separate them into 1). heatmap-based and 2). regression-based. general, methods achieve higher accuracy but are subject various heuristic designs (not end-to-end mostly), whereas attain relatively lower they have less intermediate non-differentiable steps. Here utilize encoder-decoder structure Transformers perform person keypoint...
Vision transformers (ViTs) have recently received explosive popularity, but the huge computational cost is still a severe issue. Since computation complexity of ViT quadratic with respect to input sequence length, mainstream paradigm for reduction reduce number tokens. Existing designs include structured spatial compression that uses progressive shrinking pyramid computations large feature maps, and unstructured token pruning dynamically drops redundant However, limitation existing lies in...
Abstract Transformers, the dominant architecture for natural language processing, have also recently attracted much attention from computational visual media researchers due to their capacity long-range representation and high performance. Transformers are sequence-to-sequence models, which use a self-attention mechanism rather than RNN sequential structure. Thus, such models can be trained in parallel represent global information. This study comprehensively surveys recent transformer works....
In this paper, we present a joint end-to-end line segment detection algorithm using Transformers that is post-processing and heuristics-guided intermediate processing (edge/junction/region detection) free. Our method, named LinE TRansformers (LETR), takes advantages of having integrated tokenized queries, self-attention mechanism, encoding-decoding strategy within by skipping standard heuristic designs for the edge element perceptual grouping processes. We equip with multi-scale...
We propose an algorithm, guided variational autoencoder (Guided-VAE), that is able to learn a controllable generative model by performing latent representation disentanglement learning. The learning objective achieved providing signal the encoding/embedding in VAE without changing its main backbone architecture, hence retaining desirable properties of VAE. design unsupervised and supervised strategy Guided-VAE observe enhanced modeling controlling capability over vanilla In strategy, we...
Problem definition : Mobile communications technologies and online platforms have enabled large-scale consumer-to-consumer (C2C) sharing of their underutilized products. This paper studies a manufacturer’s optimal entry strategy in the product-sharing market economic implications its entry. Academic/practical relevance Sharing products or services among consumers has experienced dramatic growth recent years. The impact C2C on traditional firms can be very significant. In response to product...
Aiming at the characteristics of remote sensing images such as a complex background, large number small targets, and various target scales, this paper presents image detection algorithm based on improved YOLOv8. First, in order to extract more information about targets images, we add an extra layer for backbone network; second, propose C2f-E structure Efficient Multi-Scale Attention Module (EMA) enhance network’s ability detect different sizes; lastly, Wise-IoU is used replace CIoU loss...
Product quality and product warranty coverage are two important closely related operational decisions. A longer protection period can boost sales, but it may also result in dramatically increased cost, if is poor. To investigate how these decisions interact with each other influence supply chain performance, we develop a single‐period model supplier that provides to an original equipment manufacturer, which turn sells customers. Customer demand random affected by the length of period....
Visual geo-localization remains a challenging task due to variations in the appearance and perspective among captured images. This paper introduces an efficient TransVLAD module, which aggregates attention-based feature maps into discriminative compact global descriptor. Unlike existing methods that generate using only convolutional neural networks (CNNs), we propose sparse transformer encode dependencies compute maps, effectively reduces visual ambiguities occurs large-scale problems. A...
The outbreak of the toxic capsule crisis during April 2012 aroused widespread public concern about risk chromium‐contaminated capsules and drug safety in China. In this article, we develop a conceptual model to investigate perceptions pharmaceutical behavioral responses relationship between associated factors these two variables. An online survey was conducted test model, including questions on measures perceived efficacy countermeasures, trust State FDA (Food Drug Administration),...
Biometrics authentication has become increasingly popular due to its security and convenience; however, traditional biometrics are becoming less desirable in scenarios such as new mobile devices, Virtual Reality, Smart Vehicles. For example, while face is widely used, it suffers from significant privacy concerns. The collection of complete facial data makes for privacy-sensitive applications. Lip authentication, on the other hand, emerged a promising method. However, existing lip-based...
The accurate and prompt diagnosis of infections is essential for improving patient outcomes preventing bacterial drug resistance. Host gene expression profiling as an approach to infection holds great potential in assisting early infection. To improve the precision diagnosis, we developed InfectDiagno, a rank-based ensemble machine learning algorithm via host patterns. Eleven data sets were used training method development, InfectDiagno was optimized by multi-cohort samples. Nine independent...
Controllable 3D scene generation has extensive applications in virtual reality and interior design, where the generated scenes should exhibit high levels of realism controllability terms geometry. Scene graphs provide a suitable data representation that facilitates these applications. However, current graph-based methods for are constrained to text-based inputs insufficient adaptability flexible user inputs, hindering ability precisely control object To address this issue, we propose...
Controllable 3D scene generation has extensive applications in virtual reality and interior design, where the generated scenes should exhibit high levels of realism controllability terms geometry. Scene graphs provide a suitable data representation that facilitates these applications. However, current graph-based methods for are constrained to text-based inputs insufficient adaptability flexible user inputs, hindering ability precisely control object To address this issue, we propose...