Crowdsourced Annotation System for Transaction Clarity
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Users face difficulties in understanding ambiguous merchant identifiers in financial transaction histories, and there is a need for a system to crowdsource annotations for transactions, automatically recognize recurring expenses, and detect fraudulent transactions.
Innovation Solution
A system that includes a crowdsourcing annotation database shared among users, where processors receive transaction data, retrieve relevant annotations, and dynamically update the database with user-provided annotations, enabling automatic annotation of new transactions, detection of fraudulent transactions, and identification of recurring expenses based on check images and text recognition.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If a system stores and retrieves crowdsourced annotations for transactions, then transaction clarity and user understanding are improved, but system complexity and data management requirements increase
Solution Approach 1:
The crowdsourced annotation system serves multiple functions: it stores annotations, retrieves relevant annotations for display, allows user contributions, and dynamically updates the database. This multi-functional approach consolidates what would otherwise require separate systems into a unified annotation management platform.
Solution Approach 2:
The system introduces an intermediary annotation layer between the raw transaction data and the user. Instead of users directly interpreting ambiguous merchant identifiers, the system mediates by providing crowdsourced annotations that bridge the gap between raw data and user understanding.
2Loss of time
If the system automatically retrieves and displays crowdsourced annotations for new transactions, then transaction recognition speed is improved, but processing time and computational resources increase
Solution Approach 1:
The system performs preliminary actions by pre-storing and organizing crowdsourced annotations in the database before they are needed. When a new transaction arrives, the system can quickly retrieve relevant annotations without needing to create them on-demand, significantly reducing processing time.
Solution Approach 2:
The system uses copying by retrieving and displaying existing crowdsourced annotations for new transactions instead of creating new annotations from scratch. This allows rapid replication of proven annotation patterns across multiple transactions, improving both speed and consistency.
3Reliability
If the system allows users to contribute and share annotations dynamically, then annotation accuracy and reliability are improved, but data validation and security requirements increase
Solution Approach 1:
The system implements feedback mechanisms where user-contributed annotations are dynamically updated in the database and become available for future transactions. This creates a continuous feedback loop where annotations are refined and improved over time based on actual usage and user contributions, enhancing reliability.
Solution Approach 2:
The system enables users to self-serve by allowing them to contribute their own annotations and share them with the community. This user-generated content approach leverages collective knowledge while reducing the burden on the system to pre-validate all annotations, though it does require implementation of appropriate validation and security measures.
Data Source
AI summary
A system for crowdsourcing annotations for transactions includes a crowdsourcing annotation database and a processor. The database stores crowdsourced annotations associated with merchants. The crowdsourced annotations are shared among, and contributed by, users of a community. The processor receives transaction data for a transaction by a user with a merchant. Relevant crowdsourced annotations associated with the merchant are retrieved from the database and sent to the user to enable to the user to annotate the transaction. The user provides an annotation for the transaction. The system dynamically updates the database based on the annotation provided by the user. In another aspect a transaction prediction system is disclosed. The system receives transaction data and identifies text on a check image associated with the transaction data. The system identifies a recurring expense and associated expense frequency, and may generate an expense warning or suggestion to execute a check to pay the expense.