Author: Olaf Kopp
Reading time: 6 Minutes

Providing search results based on a compositional query

Topics: , , ,

Rate this post

The patent describes a sophisticated method for improving search results by deeply understanding and processing compositional queries through the use of knowledge graphs and detailed attribute comparison. This advancement could significantly impact how search engines interpret and fulfill complex search queries, marking a notable improvement in search technology’s ability to cater to nuanced user needs.

  • Patent ID: US20240078274A1
  • Assignee: Google LLC (Mountain View, CA, US)
  • Inventors:
    • Jinyu Lou (Shanghai, CN)
    • Ying Chai (Shanghai, CN)
    • Chen Ding (Redwood City, CA, US)
    • Lijie Chen (Shanghai, CN)
    • Liang Hu (Hubei, CN)
    • Kejia Liu (Shanghai, CN)
    • Weibin Pan (Shanghai, CN)
    • Yanlai Huang (Shanghai, CN)
    • David Francois Huynh (San Francisco, CA, US)
  • Attorney/Agent: BRAKE HUGHES BELLERMANN LLP (Middletown, MD, US)
  • Publication Date: 03/07/2024
  • Filing Date: 09/18/2023
  • Countries: US, China, WIPO


Based on the background information provided, this Google patent application seems to be addressing the improvement of search engine capabilities, particularly in handling compositional queries. Let’s break down the key insights and implications of this innovation:

Background Understanding

  • Core Problem: Traditional search engines are adept at handling queries based on single, fixed criteria, typically either spatial (location-based) or temporal (time-based). Examples include finding a Starbucks near a specific airport or listing films made during a particular historical period. However, these queries do not always capture the complexity of user intentions that involve multiple criteria or the composition of different types of information.
  • Innovation Goal: The aim appears to be enhancing search engine algorithms to better interpret and respond to compositional queries. These are queries that combine multiple criteria, such as location and time, to provide more relevant and specific search results.

This patent was first signed by Google in 2012 and republished in an updated form in May 2021. It runs until 2035. The patent describes the process for returning search results to a search query with at least two related entities. Such a search query could be, for example, “Japanese restaurant near a bank”. The search intent behind this search query could be that the user would like to visit a Japanese restaurant and stop by a bank before or after.

Another search query could be “banks that went bankrupt during the economic crisis”. In contrast to the previous search query, this is not about a location but a temporal relationship between the entities.

The search queries can also be more complex and contain more than two entities.

Overview of the Method

  • Determining Entities and Relationships: The method involves using processors to analyze a compositional query and determine the types of entities involved and their relationships. For instance, in a query like “tech conferences in Europe in 2023,” the entity types could be “tech conferences” and “Europe,” with “2023” defining a temporal relationship.
  • Utilizing a Knowledge Graph: The method includes identifying relevant nodes within a knowledge graph. These nodes represent the entities recognized from the query. A knowledge graph is a structured way of representing data, where nodes represent entities (such as objects, locations, events), and edges represent the relationships between them.
  • Attribute Value Determination: For each identified entity, the method determines attribute values that correspond to the defined relationship in the query. Using the previous example, this could involve identifying all tech conferences (first entity type) and their locations in Europe (second entity type), along with their dates (relationship attribute).
  • Comparing Attribute Values: The method compares these attribute values across the first and second entity types to identify matches that fulfill the compositional criteria of the query.
  • Resultant Entity References: Based on this comparison, the method determines resultant entity references that best match the query’s criteria. This results in a refined set of search results that are more accurately aligned with the user’s intent.

Such complex search queries can only be answered by data recorded in a Knowledge Graph. An infinite number of references can be established between a wide variety of entities via the relationship information described by the edges. A classic tabular database such as SQL cannot answer such search queries in a meaningful way.

The detailed description provides real-world examples to illustrate the application of the methodology for processing compositional queries. Here are the examples mentioned:

  1. Spatial Relationship Query: The query example “[American Banks close to Japanese restaurants]” is used to demonstrate a scenario where a user is interested in finding American banks that are located near Japanese restaurants. The query does not specify a particular location but seeks to identify entities (banks and restaurants) based on their spatial relationship. This example illustrates how the system can handle queries that involve spatial proximity without explicit location details.
  2. Temporal Relationship Query: Another example provided is “[Companies that went bankrupt during an economic crisis],” which represents a query seeking companies that filed for bankruptcy within the timeframe of an economic crisis. This example shows the system’s ability to process queries that involve a temporal relationship between entities, in this case, companies and economic crises.

These examples are used to explain how the system recognizes and processes different types of entity references and their relationships, whether spatial or temporal. They serve to demonstrate the practical application of the system’s methodology in real-world scenarios, showcasing its ability to understand and respond to complex queries that involve multiple entities and relationships.

Implications for SEO

Understanding of Entity Relationships

  • Entity Recognition and Optimization: The patent highlights the importance of recognizing entities and their types (e.g., people, places, ideas) within web content. For SEO, this underscores the necessity to structure content in a way that clearly identifies and relates entities. This can be achieved through the use of schema markup and structured data, helping search engines understand the context and relationships between entities within content.
  • Rich Snippets and Knowledge Graphs: By optimizing for entities and their relationships, there’s potential for enhanced visibility through rich snippets and knowledge graph entries. Providing clear, structured information about entities can increase the likelihood of content being featured in these prominent search results.

Search Intent and Contextual Relevance

  • Focus on User Intent: The patent illustrates the importance of understanding the implicit intent behind compositional queries, such as spatial or temporal relationships. SEO strategies should focus on creating content that not only targets keywords but also addresses the underlying questions or intentions that users might have.
  • Contextual Content Creation: Developing content that explores the relationships between entities in depth can improve relevance in search queries that involve complex relationships. This means going beyond single-topic pages to create content that addresses the wider context of the entities involved.

Local SEO and Temporal Content

  • Enhanced Local SEO: For businesses, the patent’s approach to handling spatial relationships in queries highlights the importance of optimizing for local search. This includes ensuring that business listings are accurate and detailed, and creating content that establishes relationships between the business and its local area or other local entities.
  • Temporal Content Strategy: For queries that involve a temporal aspect, there’s an opportunity to create content that remains relevant or becomes relevant at specific times. This could involve covering events, trends, or topics that have a clear temporal relationship, ensuring that content is timely and can be surfaced in relevant searches.

Advanced SEO Strategies

  • Data Structuring and Interlinking: SEO strategies should consider how to structure data and interlink content in a way that highlights the relationships between different pieces of content. This can help search engines understand the broader context of the website’s content and how it relates to specific entities and their attributes.
  • Predictive and Proactive Content Creation: By analyzing common entity relationships within a niche, SEOs can predict potential compositional queries users might make. Creating content that proactively addresses these queries can position a website as a valuable resource, improving its visibility in search results.


Content from the blog

LLMO: How do you optimize for the answers of generative AI systems?

As more and more people prefer to ask ChatGPT rather than Google when searching for read more

What is the Google Knowledge Vault? How it works?

The Google Knowledge Vault was a project by Google that aimed to create an extensive read more

What is BM25?

BM25 is a popular ranking function used in information retrieval systems to estimate the relevance read more

The dimensions of the Google ranking

The ranking factors at Google have become more and more multidimensional and diverse over the read more

Interesting Google patents for search and SEO in 2024

In this article I would like to contribute to archiving well-founded knowledge from Google patents read more

What is the Google Shopping Graph and how does it work?

The Google Shopping Graph is an advanced, dynamic data structure developed by Google to enhance read more