Generative Summaries For Search Outcomes

Google has a published patent named “Generative summaries for search results” which is believed to be the patent behind the Search Generative Experience launch we saw earlier this year. This patent was filed on March 20, 2023 and approved on September 26, 2023 under the patent ID US11769017B1.

Juan Gonzalez Villa posted a thread on X breaking it down, which I will embed below so you can read it.

The abstract reads:

At least selectively utilizing a large language model (LLM) in generating a natural language (NL) based summary to be rendered in response to a query. In some implementations, in generating the NL based summary additional content is processed using the LLM. The additional content is in addition to query content of the query itself and, in generating the NL based summary, can be processed using the LLM and along with the query content—or even independent of the query content. Processing the additional content can, for example, mitigate occurrences of the NL based summary including inaccuracies and/or can mitigate occurrences of the NL based summary being over-specified and/or under-specified.

Here are Juan’s posts:

Before we start: there are currently no other patents assigned to Google mentioning generative AI techniques and search.

There might be other patents on the same topic still pending, but I believe this document is highly relevant to Google SGE as it works right now. pic.twitter.com/L4Ly5Ows2c

— Juan González Villa (@seostratega) November 29, 2023

1. Receiving a query associated with a client device. The query can be explicitly entered by a user or automatically generated based on context.

2. Selecting a set of search result documents (SRDs) that are responsive to the query and related or recent queries.

— Juan González Villa (@seostratega) November 29, 2023

4. Processing the SRD content snippets using an LLM to generate LLM output. An optional summarization prompt can also be included.

5. Generating a natural language summary using the LLM output. This leverages robustness of the LLM while constraining it to the SRD content.

— Juan González Villa (@seostratega) November 29, 2023

This checks out with something stated in the document “An Overview of SGE”, made public by Google around SGE’s launch, although we didn’t have any more details.

The patent now provides plenty of detail around how and why several models are available and can be used: pic.twitter.com/U035N35Bxl

— Juan González Villa (@seostratega) November 29, 2023

3. The selection can be based on:

• Processing the query with a classifier to predict best model(s)
• Detecting certain terms in the query to indicate suitable model(s)
• Considering search result types/content to determine appropriate model(s)

— Juan González Villa (@seostratega) November 29, 2023

So by dynamically selecting from multiple candidate generative models, the invention aims to optimize accuracy and efficiency by choosing the most suited model(s) for any given query.

The patent also provide some details on how the links to sources work:

— Juan González Villa (@seostratega) November 29, 2023

1. “A portion, of a visually rendered NL summary, that is supported by a first SRD can be selectable (and optionally underlined, highlighted, and/or otherwise annotated).

A selection of the portion can result in navigating to a link corresponding to the first SRD.”

— Juan González Villa (@seostratega) November 29, 2023

3. The links can be general links to the SRDs or specific anchor links to the portions that provide the verification.

4. The linkified portions can be determined based on comparing the summary content to SRD content using encoder models to identify verified portions. pic.twitter.com/zlJFFO2c6Q

— Juan González Villa (@seostratega) November 29, 2023

Here’s how confidence works, according to the patent:

1. Confidence measures can be generated for portions of the summary or for the summary as a whole.

The confidence measures are then used to determine which confidence annotation from a set of candidates should be applied.

— Juan González Villa (@seostratega) November 29, 2023

3. A textual “high confidence”, “medium confidence”, or “low confidence” annotation can be annotated for the NL based summary as a whole.

Each of the portions of the NL based summary can be annotated with a corresponding color that reflects a degree of confidence in that portion

— Juan González Villa (@seostratega) November 29, 2023

Thanks for reading so far.

Here’s the link to the patent: https://t.co/AtGMiWqzG9

Now, I’ll explain how I found out about this patent, and another interesting thing:

What Google SGE and Featured Snippets have in common… 🤓⤵

— Juan González Villa (@seostratega) November 29, 2023

According to his own career summary, he worked in SGE at Google between Nov 2022 and Sep 2023. He is now a Google Fellow, the highest rung on the ladder for Google engineers, which they reach thanks to “consistently outstanding accomplishments”. pic.twitter.com/aZmpfkajmI

— Juan González Villa (@seostratega) November 29, 2023

Their careers afterwards were very similar: they both went to Apple, where they worked in Search and Siri, and came back to Google at the end of 2022 to work in SGE.

So… how did I find the patent? I decided to look for patents by any of these two engineers and there it was. 💡 pic.twitter.com/1uUPi9myoj

— Juan González Villa (@seostratega) November 29, 2023

Their careers afterwards were very similar: they both went to Apple, where they worked in Search and Siri, and came back to Google at the end of 2022 to work in SGE.

So… how did I find the patent? I decided to look for patents by any of these two engineers and there it was. 💡 pic.twitter.com/1uUPi9myoj

— Juan González Villa (@seostratega) November 29, 2023

If you enjoyed this thread, any feedback, likes and/or RTs are appreciated.

Also, here’s a post version of this thread:https://t.co/JSt9qNcVTr

— Juan González Villa (@seostratega) November 29, 2023

Nice write up Juan!

Forum discussion at X.