Crowdsourced Email Filtering via Community Voting and Penalty Box
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing spam detection methods rely solely on keyword recognition, failing to distinguish between spam and unwanted, unsolicited emails, and lack a mechanism to effectively filter out annoying emails based on community recognition.
Innovation Solution
A crowdsourced email filtering system that allows community members to vote on unwanted emails, processing user input to identify and penalize senders through a penalty box mechanism, ensuring future emails from penalized senders are filtered out until better behavior is exhibited.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If keyword-based spam detection is used, then simple implementation is achieved, but accuracy in distinguishing spam from unwanted emails deteriorates
Solution Approach 1:
The system enables users to actively participate in spam detection by voting on emails they receive. Each user's judgment contributes to the collective filtering capability, making the system self-improving through community input rather than relying solely on pre-programmed keywords
Solution Approach 2:
The system implements a feedback loop where users vote on unwanted emails, the results are aggregated and stored, and this feedback is used to automatically filter future emails from the same senders. This continuous feedback mechanism improves detection accuracy over time
2Device complexity
If rigid keyword-based filtering is used, then system complexity is reduced, but adaptability to community-defined unwanted emails deteriorates
Solution Approach 1:
The filtering criteria dynamically adapt based on community voting patterns. Instead of static keywords, the system learns from user behavior and automatically adjusts its filtering rules to match community-defined unwanted email patterns
Solution Approach 2:
The system automatically updates its filtering logic through aggregated user votes without requiring manual reconfiguration. The community collectively defines what constitutes unwanted email, and the system self-adjusts to reflect these evolving preferences
3Measurement precision
If community voting mechanism is added, then email filtering accuracy is improved, but system complexity increases
Solution Approach 1:
Users interact with the system through simple voting actions on emails they receive, without needing to understand or configure the underlying complex aggregation and filtering algorithms. The complexity is hidden while providing accurate filtering results
Solution Approach 2:
The system introduces an intermediary layer between raw user judgments and final filtering decisions. This intermediary aggregates individual votes, applies thresholds, and translates community sentiment into automated filtering actions, managing complexity centrally
4Reliability
If penalty box mechanism is implemented, then sender penalization effectiveness is improved, but processing time increases
Solution Approach 1:
The system performs preliminary actions by pre-calculating and storing penalty thresholds and sender information in advance. When an email arrives, the system can quickly match it against pre-computed criteria rather than performing complex real-time analysis
Solution Approach 2:
For emails from penalized senders, the system skips detailed analysis and directly applies the penalty filter. Once a sender is identified as penalized through the voting mechanism, subsequent emails from that sender are rapidly rejected without re-evaluating their content
Data Source
AI summary
Disclosed herein are system, method, and computer program product embodiments for crowdsourcing e-mail filtering. An embodiment operates by receiving an input from a community member indicating that an e-mail received by the community member is unwanted, counting a number of inputs received from community members indicating that e-mail is unwanted based on one or more criteria, flagging future e-mails based on the criteria when the number of inputs counted reaches a threshold for penalizing an originator or a group associated with the originator of the unwanted e-mail, and denying delivery of future e-mails from the originator or the group associated with the originator until a penalty period expires.