In product search, the retrieval of candidate products before re-ranking is more critical and challenging than other search like web search, especially for tail queries, which have a complex and specific search intent. In this paper, we present a hybrid system for e-commerce search deployed at Walmart that combines traditional inverted index and embedding-based neural retrieval to better answer user tail queries. Our system significantly improved the relevance of the search engine, measured by both offline and online evaluations. The improvements were achieved through a combination of different approaches. We present a new technique to train the neural model at scale. and describe how the system was deployed in production with little impact on response time. We highlight multiple learnings and practical tricks that were used in the deployment of this system.

High relevance of retrieved and re-ranked items to the search query is the cornerstone of successful product search, yet measuring relevance of items to queries is one of the most challenging tasks in product information retrieval, and quality of product search is highly influenced by the precision and scale of available relevance-labelled data. In this paper, we present an array of techniques for leveraging Large Language Models (LLMs) for automating the relevance judgment of query-item pairs (QIPs) at scale. Using a unique dataset of multi-million QIPs, annotated by human evaluators, we test and optimize hyper parameters for finetuning billion-parameter LLMs with and without Low Rank Adaption (LoRA), as well as various modes of item attribute concatenation and prompting in LLM finetuning, and consider trade offs in item attribute inclusion for quality of relevance predictions. We demonstrate considerable improvement over baselines of prior generations of LLMs, as well as off-the-shelf models, towards relevance annotations on par with the human relevance evaluators. Our findings have immediate implications for the growing field of relevance judgment automation in product search.

At TREC 2023, we hosted the first TREC Product Search Track, looking to create a reusable general benchmark for evaluating the performance of retrieval methods in the product search domain. We focus on providing a benchmark similar in scale and format to NQ Kwiatkowski et al. [2019], or the Deep Learning Track Craswell et al. [2021] but focused on product search. In providing a simple-to-use dataset, we believe broad experimentation using popular retrieval libraries Lin et al. [2021] Gao et al. [2022] can lead to broad improvements in retrieval performance. In this first year of the track, we created a novel collection based on the ESCI Product Re-ranking dataset Reddy et al. [2022], sampled novel queries, created enriched metadata in the form of additional text and images along with seeded evaluation results with a broad range of baseline runs to aid in collection reusability and to allow iteration and experimentation on the use of additional context. Unlike previous product search corpora, the Product Search Track is multi-modal and has a large enough scale to explore the usage of neural retrieval methods.

In this paper, we consider the problem of improving the inference latency of language model-based dense retrieval systems by introducing structural compression and model size asymmetry between the context and query encoders. First, we investigate the impact of pre and post-training compression on the MSMARCO, Natural Questions, TriviaQA, SQUAD, and SCIFACT, finding that asymmetry in the dual encoders in dense retrieval can lead to improved inference efficiency. Knowing this, we introduce Kullback Leibler Alignment of Embeddings (KALE), an efficient and accurate method for increasing the inference efficiency of dense retrieval methods by pruning and aligning the query encoder after training. Specifically, KALE extends traditional Knowledge Distillation after bi-encoder training, allowing for effective query encoder compression without full retraining or index generation. Using KALE and asymmetric training, we can generate models which exceed the performance of DistilBERT despite having 3x faster inference.

The success of contextual word representations and advances in neural information retrieval have made dense vector-based retrieval a standard approach for passage and document ranking. While effective and efficient, dual-encoders are brittle to variations in query distributions and noisy queries. Data augmentation can make models more robust but introduces overhead to training set generation and requires retraining and index regeneration. We present Contrastive Alignment POst Training (CAPOT), a highly efficient finetuning method that improves model robustness without requiring index regeneration, the training set optimization, or alteration. CAPOT enables robust retrieval by freezing the document encoder while the query encoder learns to align noisy queries with their unaltered root. We evaluate CAPOT noisy variants of MSMARCO, Natural Questions, and Trivia QA passage retrieval, finding CAPOT has a similar impact as data augmentation with none of its overhead.

In e-commerce, product content, especially product images have a significant influence on a customer's journey from product discovery to evaluation and finally, purchase decision. Since many e-commerce retailers sell items from other third-party marketplace sellers besides their own, the content published by both internal and external content creators needs to be monitored and enriched, wherever possible. Despite guidelines and warnings, product listings that contain offensive and non-compliant images continue to enter catalogs. Offensive and non-compliant content can include a wide range of objects, logos, and banners conveying violent, sexually explicit, racist, or promotional messages. Such images can severely damage the customer experience, lead to legal issues, and erode the company brand. In this paper, we present a machine learning driven offensive and non-compliant image detection system for extremely large e-commerce catalogs. This system proactively detects and removes such content before they are published to the customer-facing website. This paper delves into the unique challenges of applying machine learning to real-world data from retail domain with hundreds of millions of product images. We demonstrate how we resolve the issue of non-compliant content that appears across tens of thousands of product categories. We also describe how we deal with the sheer variety in which each single non-compliant scenario appears. This paper showcases a number of practical yet unique approaches such as representative training data creation that are critical to solve an extremely rarely occurring problem. In summary, our system combines state-of-the-art image classification and object detection techniques, and fine tunes them with internal data to develop a solution customized for a massive, diverse, and constantly evolving product catalog.

In e-commerce, content quality of the product catalog plays a key role in delivering a satisfactory experience to the customers. In particular, visual content such as product images influences customers' engagement and purchase decisions. With the rapid growth of e-commerce and the advent of artificial intelligence, traditional content management systems are giving way to automated scalable systems. In this paper, we present a machine learning driven visual content management system for extremely large e-commerce catalogs. For a given product, the system aggregates images from various suppliers, understands and analyzes them to produce a superior image set with optimal image count and quality, and arranges them in an order tailored to the demands of the customers. The system makes use of an array of technologies, ranging from deep learning to traditional computer vision, at different stages of analysis. In this paper, we outline how the system works and discuss the unique challenges related to applying machine learning techniques to real-world data from e-commerce domain. We emphasize how we tune state-of-the-art image classification techniques to develop solutions custom made for a massive, diverse, and constantly evolving product catalog. We also provide the details of how we measure the system's impact on various customer engagement metrics.

Classifying products precisely and efficiently is a major challenge in modern e-commerce. The high traffic of new products uploaded daily and the dynamic nature of the categories raise the need for machine learning models that can reduce the cost and time of human editors. In this paper, we propose a decision level fusion approach for multi-modal product classification based on text and image neural network classifiers. We train input specific state-of-the-art deep neural networks for each input source, show the potential of forging them together into a multi-modal architecture and train a novel policy network that learns to choose between them. Finally, we demonstrate that our multi-modal network improves classification accuracy over both networks on a real-world large-scale product classification dataset that we collected from Walmart.com. While we focus on image-text fusion that characterizes e-commerce businesses, our algorithms can be easily applied to other modalities such as audio, video, physical sensors, etc.

Fisher linear discriminant analysis (LDA) can be sensitive to the problem data. Robust Fisher LDA can systematically alleviate the sensitivity problem by explicitly incorporating a model of data uncertainty in a classification problem and optimizing for the worst-case scenario under this model. The main contribution of this paper is show that with general convex uncertainty models on the problem data, robust Fisher LDA can be carried out using convex optimization. For a certain type of product form uncertainty model, robust Fisher LDA can be carried out at a cost comparable to standard Fisher LDA. The method is demonstrated with some numerical examples. Finally, we show how to extend these results to robust kernel Fisher discriminant analysis, i.e., robust Fisher LDA in a high dimensional feature space.

Fisher linear discriminant analysis (LDA) can be sensitive to the problem data.