Predicting Latent Structured Intents from Shopping Queries
Topics: Ranking, Search Intent, Search Query Processing, Shopping
The paper titled “Predicting Latent Structured Intents from Shopping Queries” presents a novel framework designed to infer and map user intents from unstructured shopping queries to structured attributes using a hybrid model. This model combines Long Short-Term Memory (LSTM) networks with autoencoders to jointly learn from user behaviors and product metadata. The approach aims to improve the quality of search results on Google Shopping by understanding and predicting user intents accurately.
The paper reminds me of my approaches I introduce in the article about Shopping Graph Optimization at Search Engine Land. Here some citations from this article:
“Large language models (LLMs) learn based on the frequency of co-occurrences that occur or, in the context of ecommerce, from co-mentions of attributes with the respective product.”
“The frequency of the attributes requested in prompts and search queries determines which attributes are important for a product entity.”
“When optimizing for the shopping graph, you should mention the relevant attributes in the data sources mentioned above, if possible.”
“The more the attributes associated with the respective product resemble the context specified in the prompt and the attributes derived from the LLM, the more likely the products will be mentioned in a response from the generative AI.”
- Identify and understand user and product-relevant attributes: Long-tail analysis of search queries and prompts is becoming increasingly important.
- Think beyond keywords: SEOs must think in terms of concepts, entities, attributes and relationships. The time of keywords as a central focus is coming to an end.