Hasil untuk "Commerce"

A. Stephen, Olivier Toubia

S. Barnes, R. Vidgen

B. Corbitt, T. Thanasankit, Han Yi

J. Gordijn, H. Akkermans

A. Tsay, N. Agrawal

Bomil Suh, Ingoo Han

Hossein Bidgoli

Celeste See Pui Ng

Kun Zhang, Jingming Zhang, Wei Cheng et al.

In the wave of generative recommendation, we present OneMall, an end-to-end generative recommendation framework tailored for e-commerce services at Kuaishou. Our OneMall systematically unifies the e-commerce's multiple item distribution scenarios, such as Product-card, short-video and live-streaming. Specifically, it comprises three key components, aligning the entire model training pipeline to the LLM's pre-training/post-training: (1) E-commerce Semantic Tokenizer: we provide a tokenizer solution that captures both real-world semantics and business-specific item relations across different scenarios; (2) Transformer-based Architecture: we largely utilize Transformer as our model backbone, e.g., employing Query-Former for long sequence compression, Cross-Attention for multi-behavior sequence fusion, and Sparse MoE for scalable auto-regressive generation; (3) Reinforcement Learning Pipeline: we further connect retrieval and ranking models via RL, enabling the ranking model to serve as a reward signal for end-to-end policy retrieval model optimization. Extensive experiments demonstrate that OneMall achieves consistent improvements across all e-commerce scenarios: +13.01\% GMV in product-card, +15.32\% Orders in Short-Video, and +2.78\% Orders in Live-Streaming. OneMall has been deployed, serving over 400 million daily active users at Kuaishou.

Eman Alashwali, Abeer Alhuzali

In 2024, Saudi Arabia's Personal Data Protection Law (PDPL) came into force. However, little work has been done to assess its implementation. In this paper, we analyzed 100 e-commerce websites operating in Saudi Arabia against the PDPL, examining the presence of a privacy policy and, if present, the policy's declarations of four items pertaining to personal data rights and practices: 1) personal data retention period, 2) the right to request the destruction of personal data, 3) the right to request a copy of personal data, and 4) a mechanism for filing complaints. Our results show that, despite national awareness and support efforts, a significant fraction of e-commerce websites in our dataset are not fully compliant: only 31% of websites in our dataset declared all four examined items in their privacy policies. Even when privacy policies included such declarations, a considerable fraction of them failed to cover required fine-grained details. Second, the majority of top-ranked e-commerce websites in our dataset (based on search results order) and those hosted on local e-commerce hosting platforms exhibited considerably higher non-compliance rates than mid- to low-ranked websites and those not hosted on local e-commerce platforms. Third, we assessed the use of Large Language Models (LLMs) as an automated tool for privacy policy analysis to measure compliance with the PDPL. We highlight the potential of LLMs and suggest considerations to improve LLM-based automated analysis for privacy policies. Our results provide a step forward in understanding the implementation barriers to data protection laws, especially in non-Western contexts. We provide recommendations for policymakers, regulators, website owners, and developers seeking to improve data protection practices and automate compliance monitoring.

Ece Ucar, E. Ertugrul Karsak

The Human Development Index (HDI) introduced by United Nations Development Programme (UNDP) offers a unique quantitative measure that encompasses advancements in three fundamental aspects of human development: health, education, and living standards. However, focusing on only three dimensions when evaluating human development performance of countries is not adequate in today’s digital world. This study proposes a data envelopment analysis (DEA)-based composite index to provide an innovation-integrated human development performance assessment tool for countries. The novel two-stage common-weight DEA-based approach proposed in here is applied in a case study examining the performance assessment of European Union (EU) countries. The first stage of the developed methodology consists of solving the novel commonweight DEA-based approach with HDI indicators as the outputs and the Gini coefficient as the input. At the second stage, innovation-based indicators from World Bank database are used to evaluate innovation efficiency of EU countries. The composite index that yields the complete ranking of EU countries in terms of innovation-integrated human development performance is computed as the product of the efficiency scores resulting from these two stages. The rankings produced by the proposed approach are compared with the HDI rankings as well as the results obtained from various common-weight DEA-based models. First published online 27 November 2025

Kembo M. Bwana, Evelyne F. Magambo

This study compares the performance of non-life insurance in Tanzania and Kenya. Specifically, the study compares the Technical and Scale Efficiency (TSE) of non-life insurance in the two countries. It further explores the sources of technical inefficiency in non-life insurance in Tanzania and Kenya. The study employs Data Envelopment Analysis (DEA) to estimate the efficiency of the non-life insurance firms by adopting two inputs: management expenses and commission paid, while premiums written, and net investment income were used as output variables. Data were extracted from the annual reports of non-life insurance companies in Tanzania and Kenya for the years 2014-2017 (four years). The study revealed that non-life insurance firms in Tanzania were more technical and scale efficient compared to their Kenya counterparts. When technical efficiency was further decomposed into pure and scale efficiency to examine what largely caused inefficiency, it was revealed that in both Tanzania and Kenya, inefficiency is largely derived from a lack of technical efficiency, which reflects issues of innovation in the sector, inappropriate management practices, operating at sub optimal size of operation and misallocation of resources in production system. This study offers information relevant for investors and policymakers to make informed decisions in the insurance sector in both countries. It further guides insurance firms on the important inputs and proper allocation of the resources in the production system. The findings also contribute to our growing understanding of the effectiveness of non-life insurance in the insurance markets of the two nations.

Abdul Galib, Nurwahyuni Syahrir, Hasnidar Hasnidar

Main Purpose -  This study aims to reveal the role of awareness of sustainable practices, culture, and perceived behavioral control in improving tax compliance as an effort to maintain sustainable business practices for MSMEs in West Sulawesi by internalizing the Pappasang Kalindaqdaq Mandar. Method -  The research method used in this study is a mixed method with a concurrent model to analyze quantitatively and qualitatively simultaneously. Main Findings -  The results confirmed the theory of planned behavior, whereby awareness of sustainable practices and culture has a significant influence on increasing MSMEs' intention to behave in a compliant manner towards taxation, but perceived behavioral control did not have a significant influence. These findings indicate that aspects of awareness of sustainable practices and internalization of Pappasang Kalindaqdaq Mandar culture have a strong dominance in explaining MSME tax compliance in West Sulawesi. Theory and Practical Implications - The strong dominance of tax awareness and culture, but not accompanied by a significant influence on perceived behavioral control, requires further investigation. Further in-depth interviews are needed to obtain more in-depth information from tax authorities and MSMEs to uncover the actual role of control.  Novelty -  This research explores non-economic aspects from various perspectives such as awareness of sustainable practices (internal), culture (external), and perceived behavioral control (control belief) in improving tax compliance (external).

Faiza Bouzemlal, Ali Nabil Belouard

Exchange rate volatility can have socioeconomic challenges and a significant impact on export diversification of major global economies. The main objective of this article is to assess the symmetric and asymmetric effect of exchange rate volatility on Algerian export diversification. For this purpose, the autoregressive linear and nonlinear distributed lag (ARDL) model and annual data for the period 1990-2023 were used.The empirical findings using the estimation for time series data reveal that the volatility of the exchange rate has a symmetric effect on export diversification. The results revealed the presence of cointegration between the variables. The relationship between exchange rate volatility and export diversification in Algeria is positive and symmetric, which is contrary to conventional wisdom, as both currency depreciation and appreciation were found to boost diversification. Economic openness and GDP per capita significantly promote diversification, while investment and infrastructure surprisingly hinder it. Inflation also has an unexpected positive effect. The model adjusts quickly to equilibrium, though short-run results show mixed results. Diagnostic tests confirm robustness, except for serial correlation, corrected via Newey-West standard errors. This article suggests that policy makers should adopt different policies to address socio-economic challenges and keep the exchange rate stable in order to promote export diversification.

Urvashi Kishnani, Sanchari Das

E-commerce mobile applications are central to global financial transactions, making their security and privacy crucial. In this study, we analyze 92 top-grossing Android e-commerce apps (58 U.S.-based and 34 international) using MobSF, AndroBugs, and RiskInDroid. Our analysis shows widespread SSL and certificate weaknesses, with approximately 92% using unsecured HTTP connections and an average MobSF security score of 40.92/100. Over-privileged permissions were identified in 77 apps. While U.S. apps exhibited fewer manifest, code, and certificate vulnerabilities, both groups showed similar network-related issues. We advocate for the adoption of stronger, standardized, and user-focused security practices across regions.

Sijie Zhao, Jing Cheng, Yaoyao Wu et al.

Although diffusion-based image genenation has been widely explored and applied, background generation tasks in e-commerce scenarios still face significant challenges. The first challenge is to ensure that the generated products are consistent with the given product inputs while maintaining a reasonable spatial arrangement, harmonious shadows, and reflections between foreground products and backgrounds. Existing inpainting methods fail to address this due to the lack of domain-specific data. The second challenge involves the limitation of relying solely on text prompts for image control, as effective integrating visual information to achieve precise control in inpainting tasks remains underexplored. To address these challenges, we introduce DreamEcom-400K, a high-quality e-commerce dataset containing accurate product instance masks, background reference images, text prompts, and aesthetically pleasing product images. Based on this dataset, we propose DreamPainter, a novel framework that not only utilizes text prompts for control but also flexibly incorporates reference image information as an additional control signal. Extensive experiments demonstrate that our approach significantly outperforms state-of-the-art methods, maintaining high product consistency while effectively integrating both text prompt and reference image information.

Pakorn Ueareeworakul, Shuman Liu, Jinghao Feng et al.

As global e-commerce rapidly expands into emerging markets, the lack of high-quality semantic representations for low-resource languages has become a decisive bottleneck for retrieval, recommendation, and search systems. In this work, we present Compass-Embedding v4, a high-efficiency multilingual embedding framework specifically optimized for Southeast Asian (SEA) e-commerce scenarios, where data scarcity, noisy supervision, and strict production constraints jointly challenge representation learning. Compass-Embedding v4 addresses three core challenges. First, large-batch contrastive training under mixed task supervision introduces systematic false negatives that degrade semantic alignment. We propose Class-Aware Masking (CAM), a lightweight modification to the InfoNCE objective that suppresses invalid in-batch negatives and improves semantic discrimination without altering training efficiency. Second, low-resource SEA languages suffer from limited and uneven data coverage. We construct a diversified training corpus through context-grounded synthetic data generation, cross-lingual translation, and structured e-commerce data construction, enabling robust multilingual and domain-specific learning. Third, production deployment requires high-throughput inference while preserving embedding quality. We combine robustness-driven large-batch training with spherical model merging to mitigate catastrophic forgetting, and optimize inference via vLLM and FP8 quantization. Extensive evaluations across multilingual benchmarks and proprietary e-commerce tasks show that Compass-Embedding v4 achieves state-of-the-art performance on major SEA languages, significantly outperforming general-purpose embedding models in domain-specific retrieval and classification, while maintaining competitive performance on high-resource languages.

Yujing Wang, Yiren Chen, Huoran Li et al.

As global e-commerce platforms continue to expand, companies are entering new markets where they encounter cold-start challenges due to limited human labels and user behaviors. In this paper, we share our experiences in Coupang to provide a competitive cold-start performance of relevance matching for emerging e-commerce markets. Specifically, we present a Cold-Start Relevance Matching (CSRM) framework, utilizing a multilingual Large Language Model (LLM) to address three challenges: (1) activating cross-lingual transfer learning abilities of LLMs through machine translation tasks; (2) enhancing query understanding and incorporating e-commerce knowledge by retrieval-based query augmentation; (3) mitigating the impact of training label errors through a multi-round self-distillation training strategy. Our experiments demonstrate the effectiveness of CSRM-LLM and the proposed techniques, resulting in successful real-world deployment and significant online gains, with a 45.8% reduction in defect ratio and a 0.866% uplift in session purchase rate.

Jayant Sachdev, Sean D Rosario, Abhijeet Phatak et al.

Accurate query-product relevance labeling is indispensable to generate ground truth dataset for search ranking in e-commerce. Traditional approaches for annotating query-product pairs rely on human-based labeling services, which is expensive, time-consuming and prone to errors. In this work, we explore the application of Large Language Models (LLMs) to automate query-product relevance labeling for large-scale e-commerce search. We use several publicly available and proprietary LLMs for this task, and conducted experiments on two open-source datasets and an in-house e-commerce search dataset. Using prompt engineering techniques such as Chain-of-Thought (CoT) prompting, In-context Learning (ICL), and Retrieval Augmented Generation (RAG) with Maximum Marginal Relevance (MMR), we show that LLM's performance has the potential to approach human-level accuracy on this task in a fraction of the time and cost required by human-labelers, thereby suggesting that our approach is more efficient than the conventional methods. We have generated query-product relevance labels using LLMs at scale, and are using them for evaluating improvements to our search algorithms. Our work demonstrates the potential of LLMs to improve query-product relevance thus enhancing e-commerce search user experience. More importantly, this scalable alternative to human-annotation has significant implications for information retrieval domains including search and recommendation systems, where relevance scoring is crucial for optimizing the ranking of products and content to improve customer engagement and other conversion metrics.

Xian Guo, Ben Chen, Siyuan Wang et al.

Query suggestion plays a crucial role in enhancing user experience in e-commerce search systems by providing relevant query recommendations that align with users' initial input. This module helps users navigate towards personalized preference needs and reduces typing effort, thereby improving search experience. Traditional query suggestion modules usually adopt multi-stage cascading architectures, for making a well trade-off between system response time and business conversion. But they often suffer from inefficiencies and suboptimal performance due to inconsistent optimization objectives across stages. To address these, we propose OneSug, the first end-to-end generative framework for e-commerce query suggestion. OneSug incorporates a prefix2query representation enhancement module to enrich prefixes using semantically and interactively related queries to bridge content and business characteristics, an encoder-decoder generative model that unifies the query suggestion process, and a reward-weighted ranking strategy with behavior-level weights to capture fine-grained user preferences. Extensive evaluations on large-scale industry datasets demonstrate OneSug's ability for effective and efficient query suggestion. Furthermore, OneSug has been successfully deployed for the entire traffic on the e-commerce search engine in Kuaishou platform for over 1 month, with statistically significant improvements in user top click position (-9.33%), CTR (+2.01%), Order (+2.04%), and Revenue (+1.69%) over the online multi-stage strategy, showing great potential in e-commercial conversion.