Technology Category
Search and Optimization: Learning to SearchData Mining & Knowledge Management: Conversational Systems for Recommendation & RetrievalMachine Learning: Large Multimodal Models (LMMs)
Embedding based retrieval for long tail search queries in ecommerce
🤖 AI Summary
To address the poor semantic retrieval performance for long-tail e-commerce search queries—caused by sparse user interaction signals—this paper proposes a semantic product retrieval method tailored for low-frequency queries. The method comprises three key components: (1) an LLM-enhanced query signal augmentation mechanism to mitigate scarcity of labeled and interaction data; (2) a domain-adaptive dual-tower architecture integrating Transformer pretraining with multi-task contrastive learning—specifically optimizing query-query and query-product pair representations; and (3) model weight ensembling coupled with human-in-the-loop annotation to construct a high-quality evaluation dataset. In live A/B testing on a real-world e-commerce platform, the proposed method achieves a 3% lift in conversion rate over conventional lexical-matching recall baselines, demonstrating substantial improvements in both retrieval accuracy for long-tail queries and overall business impact.
Technology Category
Application Category
📝 Abstract
In this abstract we present a series of optimizations we performed on the two-tower model architecture [14], training and evaluation datasets to implement semantic product search at Best Buy. Search queries on bestbuy.com follow the pareto distribution whereby a minority of them account for most searches. This leaves us with a long tail of search queries that have low frequency of issuance. The queries in the long tail suffer from very spare interaction signals. Our current work focuses on building a model to serve the long tail queries. We present a series of optimizations we have done to this model to maximize conversion for the purpose of retrieval from the catalog. The first optimization we present is using a large language model to improve the sparsity of conversion signals. The second optimization is pretraining an off-the-shelf transformer-based model on the Best Buy catalog data. The third optimization we present is on the finetuning front. We use query-to-query pairs in addition to query-to-product pairs and combining the above strategies for finetuning the model. We also demonstrate how merging the weights of these finetuned models improves the evaluation metrics. Finally, we provide a recipe for curating an evaluation dataset for continuous monitoring of model performance with human-in-the-loop evaluation. We found that adding this recall mechanism to our current term match-based recall improved conversion by 3% in an online A/B test.
Problem
Research questions and friction points this paper is trying to address.
Improving retrieval for sparse long-tail ecommerce queries
Enhancing semantic search with LLM and transformer pretraining
Optimizing model fine-tuning using multi-pair training strategies
Innovation
Methods, ideas, or system contributions that make the work stand out.
Using LLM to enhance sparse conversion signals
Pretraining transformer model on catalog data
Fine-tuning with query-query and query-product pairs
🔎 Similar Papers
No similar papers found.
