User interface including personalized feed with dynamically generated prompts

Patent metadata

Field	Value
Publication number	US-12254039-B2
Application number	US-202318125286-A
Country	US
Kind code	B2
Filing date	Mar 23, 2023
Priority date	Mar 23, 2023
Publication date	Mar 18, 2025
Grant date	Mar 18, 2025

Abstract

The disclosed technology is generally directed to a personalized feed. In one example of the technology, selected key-value pairs from a profile associated with a user are provided. Based on a prompt that includes natural-language text instructions, the selected key-value pairs, and ranked content, a large language model is used to generate: pill prompts associated with the ranked content, such that the pill prompts are information requests that are unique and personalized to have particular relevance to the user based on selected key-value pairs, and a response to each pill prompt such that the response includes content corresponding to the requested information. A content feed is displayed to the user, including displaying selectable pills to the user as part of the displayed content feed such that each selectable pill includes a corresponding pill prompt. The response to the pill prompt that corresponds to the selection is displayed to the user.

First claim

We claim:

1. An apparatus, comprising:
a device including at least one memory having processor-executable code stored therein, and at least one processor that is adapted to execute the processor-executable code, wherein the processor-executable code includes processor-executable instructions that, in response to execution, enable the device to perform actions, including:
submitting a request via a feed service device to request ranked content that is associated with a first user;
responsive to requesting the ranked content, receiving the ranked content, at the feed service device, such that the ranked content is ranked based on a relevance of content among the ranked content to the first user;
accessing a first profile that is associated with the first user, wherein the first profile includes a plurality of key-value pairs, each key-value pair of the plurality of key-value pairs including two linked elements, the two linked elements including a key that is an identifier of the key-value pair and a value that is a corresponding value for the key;
selecting at least some of the plurality of key-value pairs from the first profile for transmission to a large language model;
generating a first prompt such that the first prompt includes natural-language text instructions for the large language model;
providing the first prompt, the plurality of selected key-value pairs, and the ranked content to the large language model as input;
using the large language model to generate a plurality of pill prompts associated with the ranked content and a response to each pill prompt of the plurality of pill prompts, such that the pill prompts are personalized information requests that are personalized to have relevance to the first user based on the plurality of selected key-value pairs, and the response to each pill prompt includes content that corresponds to the requested information;
transmitting, via the feed service device, a content feed to a user device of the first user; causing the content feed to be displayed on the user device, including causing a plurality of selectable pills to be displayed on the user device as part of the displayed content feed such that each selectable pill of the plurality of selectable pills includes a corresponding pill prompt of the plurality of pill prompts;
receiving a selection of one of the selectable pills of the plurality of selectable pills via the user device;
responsive to the selection of one of the selectable pills, causing the response to the pill prompt that corresponds to the selection to be displayed on the user device; and
using the large language model to determine logical groupings of feed items among the feed items such that the logical groupings are based on commonalities among the feed items, wherein causing the content feed to be displayed on the user device includes causing the content feed to be displayed on the user device such that logical groupings of feed items are displayed together, and wherein each logical grouping of the logical groupings has a summary that summarizes the logical grouping, the actions further including: based on the ranked content and the first profile, using the large language model to generate the summaries of the logical groupings such that the summaries are unique to the first user and such that the summaries are personalized to the first user.

2. The apparatus of claim 1 , wherein the response to the pill prompt that corresponds to the selection to be displayed to the first user is a personalized subset of the content feed.

3. The apparatus of claim 1 , wherein the content feed includes feed items, and wherein each of the feed items corresponds to separate content from among the ranked content.

4. The apparatus of claim 3 , wherein at least one of the feed items corresponds to a document, a news post, or a meeting.

5. The apparatus of claim 3 , wherein using the large language model to generate the plurality of pill prompts is accomplished such that a first pill prompt of the plurality of pill prompts is general with respect to the content feed, and such that a second pill prompt of the plurality of pill prompts is associated with a first feed item of the feed items.

6. The apparatus of claim 3 , wherein using the large language model to generate the plurality of pill prompts is accomplished such that for each feed item of the feed items, at least two pill prompts of the plurality of pill prompts pertains to that feed item, and such that at least two other pill prompts of the plurality of pill prompts are general with respect to the content feed.

7. The apparatus of claim 3 , wherein each feed item of the feed items includes a summary that summarizes the corresponding content, the actions further including: based on the ranked content and the first profile, using the large language model to generate the summaries of the feed items such the summaries are unique to the first user and such that the summaries are personalized to the first user.

8. The apparatus of claim 3 , wherein each feed item of the feed items includes a title, the actions further including: based on the ranked content and the first profile, using the large language model to generate the titles of the feed items such the titles are unique to the first user and such that the titles are personalized to the first user.

9. The apparatus of claim 3 , wherein each feed item of the feed items includes a picture that represents that feed item, the actions further including: based on the ranked content and the first profile, using the large language model to generate the pictures of the feed items such the pictures are unique to the first user and such that the pictures are personalized to the first user.

10. The apparatus of claim 1 , wherein using the large language model to determine the logical groupings is accomplished by: based on the ranked content and the first profile, using the large language model to generate commonalities among plurality of feed items such that the commonalities are personalized to the first user, and determining the logical groupings based on the personalized commonalities.

11. A method, comprising:
receiving, at a feed service device, content that is associated with a first user;
accessing a user profile that is associated with the first user, wherein the profile includes a plurality of key-value pairs, each key-value pair of the plurality of key-value pairs including two linked elements, the two linked elements including a key that is an identifier of the key-value pair and a value that is a corresponding value for the key;
selecting at least some of the plurality of key-value pairs from the profile for transmission to a large language model;
generating a first prompt such that the first prompt includes natural-language text instructions for the large language model;
providing the first prompt, the selected plurality of key-value pairs, and the content to the large language model as input;
using the large language model to generate a plurality of pill prompts associated with the content, such that the pill prompts of the plurality of pill prompts are unique and personalized to have particular relevance to the first user based on the user profile;
transmitting, via the feed service device, a content feed to a user device of the first user;
causing the content feed to be displayed to on the user device such that the content feed includes the plurality of pill prompts;
responsive to a selection of a pill prompt of the plurality of pill prompts, causing content that is associated with the selected pill prompt to be displayed on the user device, and
using the large language model to determine logical groupings of feed items among the

Assignees

• Microsoft Technology Licensing Llc

Inventors

• Amoateng Yaw Oduro
• Cannerozzi Roberta
• Grubaugh Jeremy Michael
• Kent Graham
• Troy Adam Douglas
• Cronin Aaron John
• Natvig Ola
• Grammeltvedt Åsmund
• Werpachowski Roman
• Chang Wei-Han
• Pelichet Maya Angele

Classifications

• G06F40/166

  Editing, e.g. inserting or deleting · CPC title

• G06F16/438

  Presentation of query results · CPC title

• G06T11/00

  Two-dimensional [2D] image generation · CPC title

• G06F16/24578

  using ranking · CPC title

• G06F40/40

  Processing or translation of natural language (natural language analysis G06F40/20; semantic analysis G06F40/30) · CPC title