RCS Content Filtering via ML Pre-Transmission
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current content filtering techniques fail to prevent objectionable content from being transmitted via text messages, leading to wastage of computing and networking resources in deleting such content and configuring devices to block future messages.
Innovation Solution
A Rich Communication Services (RCS) system that filters content using machine learning models to determine whether to block objectionable content in messages, conserving resources by redirecting and processing messages before they reach user devices.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If content filtering is performed after content is transmitted to user devices, then the filtering can be applied to all content types, but computing and networking resources are wasted on transmitting and then deleting objectionable content
Solution Approach 1:
The patent applies preliminary action by performing content filtering at the message center before content is transmitted to user devices. The message center intercepts messages, processes content through filtering algorithms, and blocks objectionable content before it enters the transmission network, thereby preventing waste of computing and networking resources on content that would later need to be deleted.
Solution Approach 2:
The message center acts as an intermediary between users and the content distribution network. It receives messages from senders, performs filtering operations on the content, and selectively forwards only appropriate content to recipients. This intermediary position allows the system to filter content efficiently without requiring post-transmission deletion operations.
2Loss of energy
If content filtering is performed at the message center before transmission, then computing and networking resources are conserved, but the system complexity increases
Solution Approach 1:
The message center is designed with multi-functionality, serving both as a traditional message routing hub and as a content filtering system. By integrating filtering capabilities into the existing message center infrastructure, the patent avoids creating a separate complex system while still achieving pre-transmission content filtering.
Solution Approach 2:
The message center performs self-service by autonomously evaluating message content against filtering criteria and making independent decisions about which messages to forward or block. This self-service capability reduces the need for external control systems and simplifies the overall architecture while maintaining effective content filtering.
3Adaptability or versatility
If third party database lookups are used for blocking websites and applications, then content filtering can be implemented, but text messages containing objectionable content are not prevented
Solution Approach 1:
The patent extends content filtering from traditional web-based applications to the segmented message transmission channel. By applying filtering specifically to text messages while maintaining existing filtering for websites and applications, the system addresses the gap in coverage without requiring complete redesign of the existing filtering infrastructure.
Solution Approach 2:
The system performs preliminary filtering on text message content before transmission, evaluating messages against objectionable content criteria and blocking inappropriate messages proactively. This preliminary action prevents objectionable content from being transmitted in the first place, rather than reacting after transmission occurs.
Data Source
AI summary
A rich communication services (RCS) system may receive a message associated with content. The message may be transmitted by a first user device and destined for a second user device and associated with content. The RCS system may receive subscription data associated with a user of the first user device and including information indicating whether content filtering is enabled for the user. The RCS system may determine whether content filtering is enabled for the user based on the subscription data, and may process the content, when the content filtering is enabled for the user and with a machine learning model, to determine whether to filter the content. The RCS system may perform one or more actions based on whether the content is to be filtered.