Jiaqi Xi, Raghav Saboo, Luming Chen, Martin Wang, Sudeep Das
The 'Mine and Refine' framework enhances e-commerce search retrieval by optimizing semantic text embeddings through a two-stage process involving contrastive training and policy-aligned fine-tuning.
In the world of online shopping, it's crucial for search systems to find the most relevant products for users, even when their search terms are vague or unusual. This study introduces a new approach called 'Mine and Refine' to improve search results. It works by first training a system to understand the general meaning of search queries and then fine-tuning it to better distinguish between different levels of relevance, like exact matches or acceptable substitutes. The method uses advanced techniques to ensure the system learns effectively from human feedback and real-world user interactions, leading to better search results and increased user satisfaction.
