Repurchase behavior is a primary signal in large-scale retail recommendation, particularly in categories with frequent replenishment: many items in a user's next basket were previously purchased, and their timing follows stable, item-specific cadences. Yet most next basket repurchase recommendation models represent history as a sequence of discrete basket events indexed by visit order, which cannot explicitly model elapsed calendar time or update item rankings as days pass between purchases. We present CASE (Cadence-Aware Set Encoding) for next basket repurchase recommendation, which decouples item-level cadence learning from cross-item interaction, enabling explicit calendar-time modeling while remaining production-scalable. CASE represents each item's purchase history as a calendar-time signal over a fixed horizon, applies shared multi-scale temporal convolutions to capture recurring rhythms, and uses induced set attention to model cross-item dependencies with sub-quadratic complexity, allowing efficient batch inference at scale. Across three public benchmarks and a proprietary dataset, CASE consistently improves precision, recall, and NDCG at multiple cutoffs compared to strong next basket recommendation baselines. In a production-scale evaluation with tens of millions of users and a large item catalog, CASE achieves up to 8.6% relative precision lift and 9.9% relative recall lift at top-5, showing that scalable cadence-aware modeling yields measurable gains in both benchmark and industrial settings.

A fundamental challenge of recommender systems (RS) is understanding the causal dynamics underlying users’ decision making. Most existing literature addresses this problem by using causal structures inferred from domain knowledge. However, there are numerous phenomenons where domain knowledge is insufficient, and the causal mechanisms must be learned from the feedback data. Discovering the causal mechanism from RS feedback data is both novel and challenging, since RS itself is a source of intervention that can influence both the users’ exposure and their willingness to interact. Also for this reason, most existing solutions become inappropriate since they require data collected free from any RS. In this article, we first formulate the underlying causal mechanism as a causal structural model and describe CSL4RS , a general causal structure learning framework for RS grounded in the real-world working mechanism. The essence of our approach is to acknowledge the unknown nature of RS intervention. We then derive the learning objective from our framework and utilize an augmented Lagrangian solver for efficient optimization. We conduct both simulation and real-world experiments to demonstrate how our approach compares favorably to existing solutions, together with the empirical analysis from sensitivity and ablation studies.

Marketplaces with multiple sellers have progressively evolved into viable business models in many web applications. Within this sphere, a marketplace recommendation model provides personalized suggestions regarding items and sellers that correspond with users’ preferences. However, a majority of the popular recommendation models are item-centric, often neglecting the incorporation of user preferences to sellers, thereby undermining the comprehensive utilization of the seller-related information contained in the marketplace datasets.To deal with the aforementioned limitations, this study presents a novel model for marketplace recommendations. It employs multi-task learning to jointly recommend items and sellers to users. Here, we introduce the KAROL, a model comprised of two principal modules. The first module is a Knowledge Attentive Graph Convolutional Network (KAGCN) structure based on the user-item-seller knowledge graph (KG). Specifically, relation-aware graph attention and LightGCN are employed to learn node embeddings of users, items and sellers. Sellers and items function reciprocally as knowledge bases for the bipartite graphs to transfer the knowledge between different tasks. It further employs dual losses to concurrently generate recommendations for both items and sellers. The second module is Adaptive Contrastive Learning (ACL), which involves a contrastive loss that incorporates three schemes for data augmentation: cross-relation sampling, edge-dropping and noise addition, to address knowledge sharing, structural consistency, and robustness challenges. An additional innovative facet is the incorporation of an adaptive temperature that is automatically optimized for contrastive loss without manual hyperparameter tuning. The experiment results on three datasets demonstrate that our model outperforms nine baseline models on both item and seller recommendation tasks.

Online e-commerce platforms have been extending in-store shopping, which allows users to keep the canonical online browsing and checkout experience while exploring in-store shopping. However, the growing transition between online and in-store becomes a challenge to online sequential recommender systems for future online interaction prediction due to the lack of holistic modeling of hybrid user behaviors (online & in-store). The challenges are two-fold. First, combining online & in-store user behavior data into a single data schema and supporting multiple stages in the model life cycle (pre-training, training, inference, etc.) organically needs a new data pipeline design. Second, online recommender systems, which solely relies on online user behavior sequences, must be redesigned to support online and in-store user data as input under the sequential modeling setting. To overcome the first challenge, we propose a hybrid, omnichannel data pipeline to compile online & in-store user behavior data by caching information from diverse data sources. Later, we introduce a model-agnostic encoder module to the sequential recommender system to interpret the user in-store transaction and augment the modeling capacity for better online interaction prediction given the hybrid user behavior.

Text-to-image models such as stable diffusion have opened a plethora of opportunities for generating art. Recent literature has surveyed the use of text-to-image models for enhancing the work of many creative artists. Many e-commerce platforms employ a manual process to generate the banners, which is time-consuming and has limitations of scalability. In this work, we demonstrate the use of text-to-image models for generating personalized web banners with dynamic content for online shoppers based on their interactions. The novelty in this approach lies in converting users' interaction data to meaningful prompts without human intervention. To this end, we utilize a large language model (LLM) to systematically extract a tuple of attributes from item meta-information. The attributes are then passed to a text-to-image model via prompt engineering to generate images for the banner. Our results show that the proposed approach can create high-quality personalized banners for users.

Product attribute value extraction is a pivotal component in Natural Language Processing (NLP) and the contemporary e-commerce industry. The provision of precise product attribute values is fundamental in ensuring high-quality recommendations and enhancing customer satisfaction. The recently emerging Large Language Models (LLMs) have demonstrated state of-the-art performance in numerous attribute extraction tasks, without the need for domain-specific training data. Nevertheless, varying strengths and weaknesses are exhibited by different LLMs due to the diversity in data, architectures, and hyperparameters. This variation makes them complementary to each other, with no single LLM dominating all others. Considering the diverse strengths and weaknesses of LLMs, it becomes necessary to develop an ensemble method that leverages their complementary potentials.

The outfit generation problem involves recommending a complete outfit to a user based on their interests. Existing approaches focus on recommending items based on anchor items or specific query styles but do not consider customer interests in famous characters from movie, social media, etc. In this paper, we define a new Character-based Outfit Generation (COG) problem, designed to accurately interpret character information and generate complete outfit sets according to customer specifications such as age and gender. To tackle this problem, we propose a novel framework LVA-COG that leverages Large Language Models (LLMs) to extract insights from customer interests (e.g., character information) and employ prompt engineering techniques for accurate understanding of customer preferences. Additionally, we incorporate text-to-image models to enhance the visual understanding and generation (factual or counterfactual) of cohesive outfits. Our framework integrates LLMs with text-to-image models and improves the customer’s approach to fashion by generating personalized recommendations. With experiments and case studies, we demonstrate the effectiveness of our solution from multiple dimensions.

Data processing is one of the fundamental steps in machine learning pipelines to ensure data quality. Majority of the applications consider the user-defined function (UDF) design pattern for data processing in databases. Although the UDF design pattern introduces flexibility, reusability and scalability, the increasing demand on machine learning pipelines brings three new challenges to this design pattern – not low-code, not dependency-free and not knowledge-aware. To address these challenges, we propose a new design pattern that large language models (LLMs) could work as a generic data operator (LLM-GDO) for reliable data cleansing, transformation and modeling with their human-compatible performance. In the LLM-GDO design pattern, user-defined prompts (UDPs) are used to represent the data processing logic rather than implementations with a specific programming language. LLMs can be centrally maintained so users don’t have to manage the dependencies at the run-time. Fine-tuning LLMs with domain-specific data could enhance the performance on the domain-specific tasks which makes data processing knowledge-aware. We illustrate these advantages with examples in different data processing tasks. Furthermore, we summarize the challenges and opportunities introduced by LLMs to provide a complete view of this design pattern for more discussions.

Keyword extraction is one of the core tasks in natural language processing. Classic extraction models are notorious for having a short attention span which make it hard for them to conclude relational connections among the words and sentences that are far from each other. This, in turn, makes their usage prohibitive for generating keywords that are inferred from the context of the whole text. In this paper, we explore using Large Language Models (LLMs) in generating keywords for items that are inferred from the items’ textual metadata. Our modeling framework includes several stages to fine grain the results by avoiding outputting keywords that are non-informative or sensitive and reduce hallucinations common in LLM’s. We call our LLM-based framework Theme-Aware Keyword Extraction (LLM-TAKE). We propose two variations of framework for generating extractive and abstractive themes for products in an E-commerce setting. We perform an extensive set of experiments on three real data sets and show that our modeling framework can enhance accuracy-based and diversity-based metrics when compared with benchmark models.

Recent studies on Next-basket Recommendation (NBR) have achieved much progress by leveraging Personalized Item Frequency (PIF) as one of the main features, which measures the frequency of the user’s interactions with the item. However, taking the PIF as an explicit feature incurs bias towards frequent items. Items that a user purchases frequently are assigned higher weights in PIF-based recommender system and appear more frequently in the personalized recommendation list. As a result, the system will lose the fairness and balance between items that the user frequently purchases and items that the user never purchases. We refer to this systematic bias on personalized recommendation lists as frequency bias, which narrows users’ browsing scope and reduces the system utility. We adopt causal inference theory to address this issue. Considering the influence of historical purchases on users’ future interests, the user and item representations can be viewed as unobserved confounders in the causal diagram. In this paper, we propose a deconfounder model named FENDER (Frequency-aware Deconfounder for Next-basket Recommendation) to mitigate the frequency bias. With the deconfounder theory and the causal diagram we propose, FENDER decomposes PIF with a neural tensor layer to obtain substitute confounders for users and items. Then, FENDER performs unbiased recommendations considering the effect of these substitute confounders. Experimental results demonstrate that FENDER has derived diverse and fair results compared to ten baseline models on three datasets while achieving competitive performance. Further experiments illustrate how FENDER balances users’ historical purchases and potential interests.

In e-Commerce Industry, customers’ purchase decision of an item are usually influenced by the reference price of that item, which is implied within the context of the items (e.g. prices of an item set from search/recommendation) or external environments (e.g. prices from another e-Commerce platform). Despite of the prevalence and influence of the reference price on customers’ behavior, existing works in Information Retrieval domain do not exploit the value of the reference price in ranking problems. In this paper, we propose a list-wise ranking model named "Prospect-Net" by incorporating the prospect theory, which is the theoretical foundation for framing the reference price. We consider the Top-K retrieval task under a product recommendation setting, and demonstrate the effectiveness of Prospect-Net to capture various forms of reference price under different scenarios. Polynomial solutions are proposed to solve the Top-K retrieval problem for some of the cases where the reference price is dependent on the recommended set of items to the user. Both offline e valuation and online experiments are performed on a real-world industrial dataset with significant performance improvement.

Recently, Graph Neural Networks (GNNs) have proven their effectiveness for recommender systems. Existing studies have applied GNNs to capture collaborative relations in the data. However, in real-world scenarios, the relations in a recommendation graph can be of various kinds. For example, two movies may be associated either by the same genre or by the same director/actor. If we use a single graph to elaborate all these relations, the graph can be too complex to process. To address this issue, we bring the idea of pre-training to process the complex graph step by step. Based on the idea of divide-and-conquer, we separate the large graph into three sub-graphs: user graph, item graph, and user-item interaction graph. Then the user and item embeddings are pre-trained from user and item graphs, respectively. To conduct pre-training, we construct the multi-relational user graph and item graph, respectively, based on their attributes.In this paper, we propose a novel Reinforced Attentive Multi-relational Graph Neural Network (RAM-GNN) to pre-train user and item embeddings on the user and item graph prior to the recommendation step. Specifically, we design a relation-level attention layer to learn the importance of different relations. Next, a Reinforced Neighbor Sampler (RNS) is applied to search the optimal filtering threshold for sampling top-k similar neighbors in the graph, which avoids the over-smoothing issue. We initialize the recommendation model with the pre-trained user/item embeddings. Finally, an aggregation-based GNN model is utilized to learn from the collaborative relations in the user-item interaction graph and provide recommendations. Our experiments demonstrate that RAM-GNN outperforms other state-of-the-art graph-based recommendation models and multi-relational graph neural networks.

Many current recommender systems mainly focus on the product-to-product recommendations and user-to-product recommendations even during the time of events rather than modeling the typical recommendations for the target event (e.g., festivals, seasonal activities, or social activities) without addressing the multiple aspects of the shopping demands for the target event. Product recommendations for the multiple aspects of the target event are usually generated by human curators who manually identify the aspects and select a list of aspect-related products (i.e., product carousel) for each aspect as recommendations. However, building a recommender system with machine learning is non-trivial due to the lack of both the ground truth of event-related aspects and the aspect-related products. To fill this gap, we define the novel problem as the event-based product carousel recommendations in e-commerce and propose an effective recommender system based on the query-click bipartite graph. We apply the iterative clustering algorithm over the query-click bipartite graph and infer the event-related aspects by the clusters of queries. The aspect-related recommendations are powered by the click-through rate of products regarding each aspect. We show through experiments that this approach effectively mines product carousels for the target event.

The complementary item recommender system (CIRS) recommends the complementary items for a given query item. Existing CIRS models consider the item co-purchase signal as a proxy of the complementary relationship, due to the lack of human-curated labels from the huge transaction records. These methods represent items in a complementary embedding space and model the complementary relationship as a point estimation of the similarity between items vectors. However, co-purchased items are not necessarily complementary to each other. For example, customers may frequently purchase bananas and bottle water within the same transaction, but these two items are not complementary. Hence, using co-purchase signals directly as labels will aggravate the model performance. On the other hand, model evaluation will not be trustworthy if the labels for evaluation are not reflecting the true complementary relatedness. To address the above challenges from noisy labeling of the co-purchase data, we model the co-purchases of two items as a Gaussian distribution, where the mean denotes the co-purchases from the complementary relatedness, and covariance denotes the co-purchases from the noise. To do so, we represent each item as a Gaussian embedding and parameterize the Gaussian distribution of co-purchases by the means and covariances from item Gaussian embedding. To reduce the impact of the noisy labels during evaluation, we propose an independence test-based method to generate a trustworthy label set with certain confidence. Our extensive experiments on both the publicly available dataset and the large-scale real-world dataset justify the effectiveness of our proposed model in complementary item recommendations compared with the state-of-the-art models.

Selecting the optimal recommender via online exploration-exploitation is catching increasing attention where the traditional A/B testing can be slow and costly, and offline evaluations are prone to the bias of history data. Finding the optimal online experiment is nontrivial since both the users and displayed recommendations carry contextual features that are informative to the reward. While the problem can be formalized via the lens of multi-armed bandits, the existing solutions are found less satisfactorily because the general methodologies do not account for the case-specific structures, particularly for the e-commerce recommendation we study. To fill in the gap, we leverage the D-optimal design from the classical statistics literature to achieve the maximum information gain during exploration, and reveal how it fits seamlessly with the modern infrastructure of online inference. To demonstrate the effectiveness of the optimal designs, we provide semi-synthetic simulation studies with published code and data for reproducibility purposes. We then use our deployment example on Walmart.com to fully illustrate the practical insights and effectiveness of the proposed methods.

Recommender systems are powerful tools for information filtering with the ever-growing amount of online data. Despite its success and wide adoption in various web applications and personalized products, many existing recommender systems still suffer from multiple drawbacks such as large amount of unobserved feedback, poor model convergence, etc. These drawbacks of existing work are mainly due to the following two reasons: first, the widely used negative sampling strategy, which treats the unlabeled entries as negative samples, is invalid in real-world settings; second, all training samples are retrieved from the discrete observations, and the underlying true distribution of the users and items is not learned.

Product embeddings have been heavily investigated in the past few years, serving as the cornerstone for a broad range of machine learning applications in e-commerce. Despite the empirical success of product embeddings, little is known on how and why they work from the theoretical standpoint. Analogous results from the natural language processing (NLP) often rely on domain-specific properties that are not transferable to the e-commerce setting, and the downstream tasks often focus on different aspects of the embeddings. We take an e-commerce-oriented view of the product embeddings and reveal a complete theoretical view from both the representation learning and the learning theory perspective. We prove that product embeddings trained by the widely-adopted skip-gram negative sampling algorithm and its variants are sufficient dimension reduction regarding a critical product relatedness measure. The generalization performance in the downstream machine learning task is controlled by the alignment between the embeddings and the product relatedness measure. Following the theoretical discoveries, we conduct exploratory experiments that supports our theoretical insights for the product embeddings.

The problem of basket recommendation (BR) is to recommend a ranking list of items to the current basket. Existing methods solve this problem by assuming the items within the same basket are correlated by one semantic relation, thus optimizing the item embeddings. However, this assumption breaks when there exist multiple intents within a basket. For example, assuming a basket contains {bread, cereal, yogurt, soap, detergent} where {bread, cereal, yogurt} are correlated through the "breakfast" intent, while {soap, detergent} are of "cleaning" intent, ignoring multiple relations among the items spoils the ability of the model to learn the embeddings. To resolve this issue, it is required to discover the intents within the basket. However, retrieving a multi-intent pattern is rather challenging, as intents are latent within the basket. Additionally, intents within the basket may also be correlated. Moreover, discovering a multi-intent pattern requires modeling high-order interactions, as the intents across different baskets are also correlated. To this end, we propose a new framework named as Multi-Intent Translation Graph Neural Network (MITGNN). MITGNN models T intents as tail entities translated from one corresponding basket embedding via T relation vectors. The relation vectors are learned through multi-head aggregators to handle user and item information. Additionally, MITGNN propagates multiple intents across our defined basket graph to learn the embeddings of users and items by aggregating neighbors. Extensive experiments on two real-world datasets prove the effectiveness of our proposed model on both transductive and inductive BR. The code1is available online.

Recommender Systems have become an integral part of online e-Commerce platforms, driving customer engagement and revenue. Most popular recommender systems attempt to learn from users' past engagement data to understand behavioral traits of users and use that to predict future behavior. In this work, we present an approach to use causal inference to learn user-attribute affinities through temporal contexts. We formulate this objective as a Probabilistic Machine Learning problem and apply a variational inference based method to estimate the model parameters. We demonstrate the performance of the proposed method on the next attribute prediction task on two real world datasets and show that it outperforms standard baseline methods.

E-commerce platforms consistently aim to provide personalized recommendations to drive user engagement, enhance overall user experience, and improve business metrics. Most e-commerce platforms contain multiple carousels on their homepage, each attempting to capture different facets of the shopping experience. Given varied user preferences, optimizing the placement of these carousels is critical for improved user satisfaction. Furthermore, items within a carousel may change dynamically based on sequential user actions, thus necessitating online ranking of carousels. In this work, we present a scalable end-to-end production system to optimally rank item-carousels in real-time on the Walmart online grocery homepage. The proposed system utilizes a novel model that captures the user's affinity for different carousels and their likelihood to interact with previously unseen items. Our system is flexible in design and is easily extendable to settings where page components need to be ranked. We provide the system architecture consisting of a model development phase and an online inference framework. To ensure low-latency, various optimizations across these stages are implemented. We conducted extensive online evaluations to benchmark against the prior experience. In production, our system resulted in an improvement in item discovery, an increase in online engagement, and a significant lift on add-to-carts (ATCs) per visitor on the homepage.

Within-basket recommendation reduces the exploration time of users, where the user's intention of the basket matters. The intent of a shopping basket can be retrieved from both user-item collaborative filtering signals and multi-item correlations. By defining a basket entity to represent the basket intent, we can model this problem as a basket-item link prediction task in the User-Basket-Item (UBI) graph. Previous work solves the problem by leveraging user-item interactions and item-item interactions simultaneously. However, collectivity and heterogeneity characteristics are hardly investigated before. Collectivity defines the semantics of each node which should be aggregated from both directly and indirectly connected neighbors. Heterogeneity comes from multi-type interactions as well as multi-type nodes in the UBI graph. To this end, we propose a new framework named BasConv, which is based on the graph convolutional neural network. Our BasConv model has three types of aggregators specifically designed for three types of nodes. They collectively learn node embeddings from both neighborhood and high-order context. Additionally, the interactive layers in the aggregators can distinguish different types of interactions. Extensive experiments on two real-world datasets prove the effectiveness of BasConv.

Recommender systems based on collaborative filtering are highly vulnerable to data poisoning attacks, where a determined attacker injects fake users with false user-item feedback, with an objective to either corrupt the recommender system or promote/demote a target set of items. Recently, differential privacy was explored as a defense technique against data poisoning attacks in the typical machine learning setting. In this paper, we study the effectiveness of differential privacy against such attacks on matrix factorization based collaborative filtering systems. Concretely, we conduct extensive experiments for evaluating robustness to injection of malicious user profiles by simulating common types of shilling attacks on real-world data and comparing the predictions of typical matrix factorization with differentially private matrix factorization.

Learning product representations that reflect complementary relationship plays a central role in e-commerce recommender system. In the absence of the product relationships graph, which existing methods rely on, there is a need to detect the complementary relationships directly from noisy and sparse customer purchase activities. Furthermore, unlike simple relationships such as similarity, complementariness is asymmetric and non-transitive. Standard usage of representation learning emphasizes on only one set of embedding, which is problematic for modelling such properties of complementariness.

In this paper, we propose a new product knowledge graph (PKG) embedding approach for learning the intrinsic product relations as product knowledge for e-commerce. We define the key entities and summarize the pivotal product relations that are critical for general e-commerce applications including marketing, advertisement, search ranking and recommendation. We first provide a comprehensive comparison between PKG and ordinary knowledge graph (KG) and then illustrate why KG embedding methods are not suitable for PKG learning. We construct a self-attention-enhanced distributed representation learning model for learning PKG embeddings from raw customer activity data in an end-to-end fashion. We design an effective multi-task learning schema to fully leverage the multi-modal e-commerce data. The ¶oincare embedding is also employed to handle complex entity structures. We use a real-world dataset from \textslgrocery.walmart.com to evaluate the performances on knowledge completion, search ranking and recommendation. The proposed approach compares favourably to baselines in knowledge completion and downstream tasks.

Conceptual relevancy - defined as how well a pair of products are related to each other - plays a significant role in online-grocery shopping behavior. When a customer goes grocery shopping, they first come up with a list of ingredients for recipes they may be interested in. A typical shopping list contains categories of products, many of them conceptually relevant to each other. However, the said user may be flexible on which exact product to purchase. For instance, a user may put `milk,' and `cheese' in their shopping list, but these may not refer to specific products such as `Great value 2% milk,' or `Kraft Singles American Slices.' Modern recommender systems, however, focus much more on how to identify specific items to recommend to customers, rather than the categories that may be relevant. Such an approach may lead the system to occasionally recommend outlier, or noise, ultimately violating conceptual relevancy. Moreover, many recommender systems ignore seasonal components; they assume that customers' shopping behavior is independent of the time of the year. However, conceptual relevancy between two products shifts over time. For instance, if a user is shopping for groceries in the middle of the winter, recommending particular products (say, barbecue-related products) may not be the best strategy even if the contextual (user's past interests, or item that the user is currently viewing) may suggest otherwise. In this paper, we introduce a novel strategy to enforce conceptual relevancy in online-grocery domain. Furthermore, recognizing that conceptual relevancy is heavily influenced by the time of the year, we propose a Bayesian-based seasonality algorithm to capture the drift in conceptual relevancy over time without having to re-train the whole model. The algorithm can be based on any of the popular approaches in recommender systems - either based on matrix factorization or that on neural networks. Through our experiments, we show that our seasonality framework can capture drifts in conceptual relevancy.

Big data computing is a process to handle large volumes of information, which typically crosses different functional units in a distributed system. Like any processes involving distributed systems, it has a concern of reliability problems, such as lossy communication links between functional units and crashed computation nodes inside functional units. The paper focuses on resolving this concern in a particular distributed system scenario where the cross-boundary network connections have a high rate of failure and the internal computation nodes are relatively reliable. We propose a pure client side protocol to achieve exactly once message processing which makes big data computing in the above scenario more reliable. Moreover, we optimize the protocol to be more efficient in resource consumption using methods such as machine learning.

In the sponsored search model, search engines are paid by businesses that are interested in displaying ads for their site alongside the search results. Businesses bid for keywords, and their ad is displayed when the keyword is queried to the search engine. An important problem in this process is keyword generation: given a business that is interested in launching a campaign, suggest keywords that are related to that campaign. We address this problem by making use of the query logs of the search engine. We identify queries re-lated to a campaign by exploiting the associations between queries and URLs as they are captured by the user’s clicks. These queries form good keyword suggestions since they cap-ture the “wisdom of the crowd ” as to what is related to a site. We formulate the problem as a semi-supervised learn-ing problem, and propose algorithms within the Markov Random Field model. We perform experiments with real query logs, and we demonstrate that our algorithms scale to large query logs and produce meaningful results.