solution Use Case
DataMiner as an AI engine
In a dynamic broadcasting environment, ensuring the quality and appropriateness of audience messages before they reach the on-air queue is crucial. To keep a close eye on this, a prominent public broadcaster used DataMiner’s capabilities to implement a tailored AI engine. This custom solution, integrated with DataMiner Automation and custom API calls, enables precise analysis of text, sentiment categorization (positive or negative), and identification of potentially offensive content.
The broadcaster’s adoption of DataMiner underscores the platform’s versatility in addressing real-world challenges. With ongoing developments, including the forthcoming DataMiner “Co-pilot” feature, users can anticipate even greater synergy between advanced AI capabilities and intuitive platform features.
USE CASE DETAILS
The foundation of this initiative lies in leveraging DataMiner's flexible API capabilities. Through a custom API call to the DataMiner System, a stream of messages, each tagged with a timestamp and unique identifier, undergoes comprehensive processing by the broadcaster's custom AI engine.
This process includes sentiment labeling and flagging of offensive content, if detected. Sentiment analysis categorizes messages as "Positive", "Negative", or triggers a "Warning" for potentially offensive content.
To simplify the process of retraining Machine Learning (ML) models, an Automation script was developed within DataMiner. This script allows users to select a dataset and choose between training either the Sentiment Detection Model or the Offensive Language Model, ensuring continuous optimization of AI-driven insights.