Personalized Theme-Based Item Collections Using Language Models
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current online concierge systems face challenges in automatically generating personalized collections of items around a coherent theme at scale, as the process is highly manual and not scalable, leading to wasted resources and late detection of trends.
Innovation Solution
Utilizing a language model and a trained computer model to generate personalized item collections based on user preferences and trends, eliminating the need for manual curation and enabling large-scale, adaptive theme-based item selection.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If manual curation process is used to create item collections, then collection quality and theme coherence can be maintained, but the process is highly manual and not scalable
Solution Approach 1:
The system enables automated collection generation where the computer model independently identifies items, determines themes, and creates collections without human intervention. The model processes user data, generates candidate items, and produces personalized collections automatically, making the system self-sufficient in the collection creation process.
Solution Approach 2:
The patent replaces the manual mechanical process of item selection and collection curation with an automated computer model. The model uses algorithmic processing to identify items, determine themes, and generate collections, substituting human manual work with computational automation that scales efficiently.
2Extent of automation
If automated query-based process is used to find items with common attributes, then scalability is improved, but it is not technically achievable to generate collections around a coherent theme
Solution Approach 1:
The system changes the parameters of automated item selection by moving beyond simple attribute-based queries to theme-based selection. The computer model incorporates thematic coherence as a selection parameter, evaluating items based on their relevance to generated themes rather than just shared attributes, enabling both automation and theme-based organization.
Solution Approach 2:
The patent introduces an intermediary theme generation process between automated item identification and collection creation. The computer model first generates candidate items, then determines coherent themes that connect these items, serving as an intermediary step that enables theme-based automated collection generation rather than direct attribute-based selection.
3Device complexity
If same collections are shown to every user, then system complexity is reduced, but valuable space is wasted on items that may not be relevant to some users
Solution Approach 1:
The system applies local quality by personalizing collections for each user based on their specific preferences, purchase history, and behavior data. Instead of uniform collections for all users, the computer model generates customized collections tailored to individual user characteristics, ensuring each user sees relevant items in their designated space.
Solution Approach 2:
The patent performs preliminary action by pre-processing user data and preferences before collection generation. The system analyzes user behavior patterns, purchase history, and preferences in advance, enabling the computer model to efficiently generate personalized collections without real-time complex computations when users view the interface.
4Loss of time
If manual trend detection is used by merchandising managers, then trend understanding can be achieved, but detection is late and requires significant employee pool
Solution Approach 1:
The system implements continuous useful action by continuously monitoring and analyzing user data, purchase patterns, and behavioral signals in real-time. The computer model continuously processes incoming data streams to detect emerging trends as they occur, rather than relying on periodic manual analysis, enabling timely and continuous trend detection.
Solution Approach 2:
The patent incorporates feedback mechanisms where the system continuously receives feedback from user interactions, purchases, and behavior patterns. This feedback loop enables the computer model to detect trends by analyzing aggregated user responses and adjust collection generation accordingly, providing timely trend detection through systematic feedback analysis.
Data Source
AI summary
An online system automatically generates a personalized collection of items around a theme. The online system generates a prompt for input into a language model, the prompt including information about a plurality of items and a text describing the theme around which the collection of items will be built. The online system requests the language model to generate, based on the prompt, a list of products eligible for building the collection of items. The online system accesses a computer model trained to identify a set of items personalized for a user of the online system. The computer model identifies, based on the list of eligible products and information about the user, the set of items for populating the collection of items. The online system causes a device of the user to display a user interface with the collection of items for inclusion into a cart of the user.