A complementary item is an item that pairs well with another item when consumed together. In the context of e-commerce, providing recommendations for complementary items is essential for both customers and stores. Current models for suggesting complementary items often rely heavily on user behavior data, such as co-purchase relationships. However, just because two items are frequently bought together does not necessarily mean they are truly complementary. Relying solely on co-purchase data may not align perfectly with the goal of making meaningful complementary recommendations. In this paper, we introduce the concept of "coherent complement recommendation", where "coherent" implies that recommended item pairs are compatible and relevant. Our approach builds upon complementary item pairs, with a focus on ensuring that recommended items are well used together and contextually relevant. To enhance the explainability and coherence of our complement recommendations, we fine-tune the Large Language Model (LLM) with coherent complement recommendation and explanation generation tasks since LLM has strong natural language explanation generation ability and multi-task fine-tuning enhances task understanding. Experimental results indicate that our model can provide more coherent complementary recommendations than existing state-of-the-art methods, and human evaluation validates that our approach achieves up to a 48% increase in the coherent rate of complement recommendations.

Understanding customer behavior is crucial for improving service quality in large-scale E-commerce. This paper proposes C-STAR, a new framework that learns compact representations from customer shopping journeys, with good versatility to fuel multiple downstream customer-centric tasks. We define the notion of shopping trajectory that encompasses customer interactions at the level of product categories, capturing the overall flow of their browsing and purchase activities. C-STAR excels at modeling both inter-trajectory distribution similarity-the structural similarities between different trajectories, and intra-trajectory semantic correlation-the semantic relationships within individual ones. This coarse-to-fine approach ensures informative trajectory embeddings for representing customers. To enhance embedding quality, we introduce a pre-training strategy that captures two intrinsic properties within the pre-training data. Extensive evaluation on large-scale industrial and public datasets demonstrates the effectiveness of C-STAR across three diverse customer-centric tasks. These tasks empower customer profiling and recommendation services for enhancing personalized shopping experiences on our E-commerce platform.

Sequential recommendation requires understanding the dynamic patterns of users' behaviors, contexts, and preferences from their historical interactions. While most research emphasizes item-level user-item interactions, they often overlook underlying shopping intentions, such as preferences for ballpoint pens or miniatures. Identifying these latent intentions is vital for enhancing shopping experiences on platforms like Amazon. Despite its significance, the area of main shopping intention detection remains under-investigated in the academic literature. To fill this gap, we introduce a graph-regularized stochastic Transformer approach, G-STO. It considers intentions as product sets and user preferences as intention composites, both modeled as stochastic Gaussian embeddings in latent space. We also employ a global intention relational graph as prior knowledge for regularization, ensuring related intentions are distributionally close. These regularized embeddings are then input into Transformer-based models to capture sequential intention transitions. On testing our model with three real-world datasets, it outperformed the baselines by 18.08% in Hit@1, 7.01% in Hit@10, and 6.11% in NDCG@10.

Sequential recommendation aims to model dynamic user behavior from historical interactions. Existing methods rely on either explicit item IDs or general textual features for sequence modeling to understand user preferences. While promising, these approaches still struggle to model cold-start items or transfer knowledge to new datasets. In this paper, we propose to model user preferences and item features as language representations that can be generalized to new items and datasets. To this end, we present a novel framework, named Recformer, which effectively learns language representations for sequential recommendation. Specifically, we propose to formulate an item as a "sentence" (word sequence) by flattening item key-value attributes described by text so that an item sequence for a user becomes a sequence of sentences. For recommendation, Recformer is trained to understand the "sentence" sequence and retrieve the next "sentence". To encode item sequences, we design a bi-directional Transformer similar to the model Longformer but with different embedding layers for sequential recommendation. For effective representation learning, we propose novel pretraining and finetuning methods which combine language understanding and recommendation tasks. Therefore, Recformer can effectively recommend the next item based on language representations. Extensive experiments conducted on six datasets demonstrate the effectiveness of Recformer for sequential recommendation, especially in low-resource and cold-start settings.

Sequential recommendation aims to model dynamic user behavior from historical interactions. Self-attentive methods have proven effective at capturing short-term dynamics and long-term preferences. Despite their success, these approaches still struggle to model sparse data, on which they struggle to learn high-quality item representations. We propose to model user dynamics from shopping intents and interacted items simultaneously. The learned intents are coarse-grained and work as prior knowledge for item recommendation. To this end, we present a coarse-to-fine self-attention framework, namely CaFe, which explicitly learns coarse-grained and fine-grained sequential dynamics. Specifically, CaFe first learns intents from coarse-grained sequences which are dense and hence provide high-quality user intent representations. Then, CaFe fuses intent representations into item encoder outputs to obtain improved item representations. Finally, we infer recommended items based on representations of items and corresponding intents. Experiments on sparse datasets show that CaFe outperforms state-of-the-art self-attentive recommenders by 44.03% [email protected] on average.

In e-commerce websites, multiple related product recommendations are usually organized into “widgets”, each given a name, as a recommendation caption, to describe the products within. These recommendation captions are usually manually crafted and generic in nature, making it difficult to attach meaningful and informative names at scale. As a result, the captions are inadequate in helping customers to better understand the connection between the multiple recommendations and make faster product discovery.

Existing recommender systems in the e-commerce domain primarily focus on generating a set of relevant items as recommendations; however, few existing systems utilize underlying item attributes as a key organizing principle in presenting recommendations to users. Mining important attributes of items from customer perspectives and presenting them along with item sets as recommendations can provide users more explainability and help them make better purchase decision. In this work, we generalize the attribute-aware item-set recommendation problem, and develop a new approach to generate sets of items (recommendations) with corresponding important attributes (explanations) that can best justify why the items are recommended to users. In particular, we propose a system that learns important attributes from historical user behavior to derive item set recommendations, so that an organized view of recommendations and their attribute-driven explanations can help users more easily understand how the recommendations relate to their preferences. Our approach is geared towards real world scenarios: we expect a solution to be scalable to billions of items, and be able to learn item and attribute relevance automatically from user behavior without human annotations. To this end, we propose a multi-step learning-based framework called Extract-Expect-Explain (EX3), which is able to adaptively select recommended items and important attributes for users. We experiment on a large-scale real-world benchmark and the results show that our model outperforms state-of-the-art baselines by an 11.35% increase on NDCG with adaptive explainability for item set recommendation.

If one customer buys a tennis racket, what are the best 3 complementary products to purchase together? 3 tennis ball packs, 3 headbands, 3 overgrips, or 1 of each respectively? Complementary product recommendation (CPR), aiming at providing product suggestions that are often bought together to serve a joint demand, forms a pivotal component of e-commerce service, however, existing methods are far from optimal. Given one product, how to recommend its complementary products of different types is the key problem we tackle in this work. We first conduct an extensive analysis to correct the inaccurate assumptions adopted by existing work to show that co-purchased products are not always complementary and further propose a new strategy to generate clean distant supervision labels for CPR modeling. Moreover, to bridge in the gap from existing work that CPR does not only need relevance modeling but also requires diversity to fulfill the whole purchase demand, we develop a deep learning framework, P-Companion to explicitly model both relevance and diversity. More specifically, given one product with its product type, P-Companion first uses an encoder-decoder network to predict multiple complementary product types, then a transfer metric learning network is developed to project the embedding of query product to each predicted complementary product type subspace and further learn the complementary relationship based on the distant supervision labels within each subspace. The whole framework can be trained from end-to-end and robust to cold-start products attributed to a novel pretrained product embedding module named Product2Vec, based on graph attention networks. Extensive offline experiments show that P-Companion outperforms state-of-the-art baselines by 7.1% increase on the [email protected] score with well-controlled diversity. Production-wise, we deploy P-Companion to provide online recommendations for over 200M products at Amazon and observe significant gains on product sales and profit.

Many stream-processing systems enforce an order on data streams during query evaluation to help unblock blocking operators and purge state from stateful operators. Such in-order processing (IOP) systems not only must enforce order on input streams, but also require that query operators preserve order. This order-preserving requirement constrains the implementation of stream systems and incurs significant performance penalties, particularly for memory consumption. Especially for high-performance, potentially distributed stream systems, the cost of enforcing order can be prohibitive. We introduce a new architecture for stream systems, out-of-order processing (OOP), that avoids ordering constraints. The OOP architecture frees stream systems from the burden of order maintenance by using explicit stream progress indicators, such as punctuation or heartbeats, to unblock and purge operators. We describe the implementation of OOP stream systems and discuss the benefits of this architecture in depth. For example, the OOP approach has proven useful for smoothing workload bursts caused by expensive end-of-window operations, which can overwhelm internal communication paths in IOP approaches. We have implemented OOP in two stream systems, Gigascope and NiagaraST. Our experimental study shows that the OOP approach can significantly outperform IOP in a number of aspects, including memory, throughput and latency.

To address increasing traffic congestion and its associated consequences, traffic managers are turning to intelligent transportation management. The latte project is extending data stream technology to handle queries that combine live streams with large data archives, motivated by needs in the Intelligent Transportation Systems (ITS) domain. In particular, we focus on queries that combine live data streams with large data archives. We demonstrate such stream-archive queries via the travel-time estimation problem. The demonstration uses the new latte system which has been developed using the NiagaraST stream processing system and the PORTAL transportation data archive.

A windowed query operator breaks a data stream into possibly overlapping subsets of data and computes a result over each. Many stream systems can evaluate window aggregate queries. However, current stream systems suffer from a lack of an explicit definition of window semantics. As a result, their implementations unnecessarily confuse window definition with physical stream properties. This confusion complicates the stream system, and even worse, can hurt performance both in terms of memory usage and execution time. To address this problem, we propose a framework for defining window semantics, which can be used to express almost all types of windows of which we are aware, and which is easily extensible to other types of windows that may occur in the future. Based on this definition, we explore a one-pass query evaluation strategy, the Window-ID (WID) approach, for various types of window aggregate queries. WID significantly reduces both required memory space and execution time for a large class of window definitions. In addition, WID can leverage punctuations to gracefully handle disorder. Our experimental study shows that WID has better execution-time performance than existing window aggregate query evaluation options that retain and reprocess tuples, and has better latency-accuracy tradeoffs for disordered input streams compared to using a fixed delay for handling disorder.