Author: Olaf Kopp
Reading time: 4 Minutes

Multi source extraction and scoring of short query answers

Topics: , , , , , ,

Rate this post

The patent aims to improve the quality of short answers provided by search engines. The patent focuses on generating short answers for search queries. It involves a training operation on a corpus of training data to train a score prediction engine. This engine is used to select and score passages from search results, determining the best short answer to display in search engine callouts.

  • Patent ID: US20230342411A1
  • Assignee: Google LLC
  • Countries: United States, China, European Patent Office, South Korea, Japan, World Intellectual Property Organization
  • Last Publishing Date: October 26, 2023
  • Inventors: Preyas Dalsukhbhai Popat, Gaurav Bhaskar Gite, John Blitzer, Jayant Madhavan, Aliaksei Severyn
  • Expiration Date: Typically, patents expire 20 years from the filing date, which would make the expiration around March 9, 2042, given the filing date of March 9, 2022.


Search engines can display quick, prominent answers to users without requiring them to click on results. This is possible because these search engines don’t rely on a curated database; instead, they provide answers to a variety of questions directly within the search results, increasing the efficiency and speed of information retrieval.


The claims of the patent focus on a method for delivering short, accurate answers directly in search results without user navigation. Key elements include:

  • Data Training: Training a score prediction engine using a dataset containing text passages and titles used previously for generating answers.
  • Selection and Scoring: When a search query is received, the system selects relevant text passages and scores them to determine their accuracy in providing the correct answer.
  • Display: The highest-scoring passage is displayed as a short answer directly in the search results, enhancing user experience by providing immediate information.
This diagram likely shows the overall architecture of the search system that incorporates the improved techniques. It could illustrate how the search engine interfaces with user devices and the Internet, showcasing the flow of data from the user query to the delivery of search results and short answers.
This diagram probably illustrates the flow of data as the system processes a search query, showing the selection of passages, the scoring process, and how the top-scoring answer is chosen and displayed.

Detailed Description

The detailed description of the patent outlines a method for enhancing search engine functionality by providing short answers directly in the search results. This method is designed to quickly deliver accurate information without the need for users to click through to different pages, which improves user experience significantly. Here are the main points:

  • Short Answer Generation: The system generates short answers by extracting passages from a top-ranked search result or from multiple relevant search results. This extraction is based on a user’s query and the relevance of the information within these passages.
  • Accuracy Scoring: An accuracy score prediction engine evaluates these passages. The score is determined by the consensus among various passages from different sources, ensuring that the displayed answer is likely accurate.
  • Contextual Validation: To improve the quality of short answers, the system uses context from other passages. This context helps to validate the answers and ensures they are not only relevant but also accurate.
  • Efficiency and Reliability: The technique emphasizes quick and reliable answers, reducing the need for users to sift through multiple search results, thus saving time and potentially reducing the load on the search engine servers.

Factors for Scoring or Ranking

  • Relevance of Passages: The system scores candidate passages on how relevant they are to the user’s search query. This relevance is likely determined by how well the content of the passage matches the query terms and the intent behind the query.
  • Consensus with Context Passages: The accuracy score for a candidate passage (potential short answer) is influenced by its consensus with other related passages. This means that if multiple sources provide similar answers or information that supports the candidate passage, its score might be higher, indicating greater reliability and accuracy.
  • Comparison with Top-Ranked Results: The methodology includes comparing the candidate passage against top-ranked search results to ensure that the information is consistent with what is considered most authoritative or credible on the subject.

Implications for SEO

  • Focus on Answer Quality and Relevance: SEO professionals should prioritize creating content that answers questions directly and succinctly. Content that is likely to serve as a source for short answer callouts should be factual, authoritative, and clearly answer specific questions. This aligns with optimizing for featured snippets in search engines.
  • Content Depth and Context: The system described uses context passages in addition to primary passages to determine the accuracy of answers. This suggests that content surrounding the answers should support and be consistent with the answers provided. Websites should include comprehensive details that reinforce the primary content, potentially improving their trustworthiness and the accuracy score assigned by search engines.
  • Semantic Consistency: The scoring system in the patent emphasizes consensus among different sources. Therefore, it’s advantageous to ensure that your content is not only accurate but also semantically in line with other reputable sources. This might involve more collaborative and comparative content research to align with the general consensus on factual information.
  • Authority and Trustworthiness: Since the system might evaluate the reliability of answers based on the source’s authority, enhancing your site’s E-A-T (Expertise, Authoritativeness, Trustworthiness) becomes crucial. This includes having clear authorship, citing authoritative sources, and maintaining a strong reputation within the industry.
  • Monitoring Competing Sources: Knowing other top-ranking sources and ensuring your content aligns or exceeds the quality and accuracy offered by these sources could be beneficial. Regular audits and updates to content will help maintain its competitiveness for being selected as a short answer.
  • Keyword and Question Targeting: Incorporating clear, search-friendly questions and their answers within your content might help in aligning with the types of queries this new system aims to respond to directly in the 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