QUT Researchers Developed an Algorithm to Help Track Abusive Content
Riley von Niessen — August 29, 2020 — Tech
References: thenextweb
A team of researchers from Australia's Queensland University of Technology have worked to develop a misogyny-detecting algorithm that can be used to filter out harmful and abusive content on Twitter.
To create the algorithm, the researchers mined 1 million Tweets from the social platform and then searched them for keywords. After this, they categorized the remaining Tweets and input the findings into a "machine learning classifier, which used the samples to create its own classification model." As the AI learning system continued to refine its vocabulary, the researchers monitored context and intent so that it could better understand if the content was meant to be humorous or otherwise non-abusive.
Although the algorithm was specifically designed to filter out misogynistic words and phrases that are ill-intentioned, it could be used similarly to filter content that is racist, homophobic, or ableist. The algorithm has not been adopted by the social platform as of yet, however the researchers hope that Twitter and other sites integrate it to help protect their users from online abuse.
Image Credit: Shutterstock
To create the algorithm, the researchers mined 1 million Tweets from the social platform and then searched them for keywords. After this, they categorized the remaining Tweets and input the findings into a "machine learning classifier, which used the samples to create its own classification model." As the AI learning system continued to refine its vocabulary, the researchers monitored context and intent so that it could better understand if the content was meant to be humorous or otherwise non-abusive.
Although the algorithm was specifically designed to filter out misogynistic words and phrases that are ill-intentioned, it could be used similarly to filter content that is racist, homophobic, or ableist. The algorithm has not been adopted by the social platform as of yet, however the researchers hope that Twitter and other sites integrate it to help protect their users from online abuse.
Image Credit: Shutterstock
Trend Themes
1. Misogyny-detecting Algorithms - The development of AI-based applications that can detect language abuse can lead to a new wave of social media safety tools.
2. AI-based Text Classification - The advancement of machine learning technology in text classification can be a disruptive innovation in content moderation.
3. Contextual Analysis Algorithm - Creating algorithms that can understand context can lead to more sophisticated applications that can recognize humor and sarcasm more accurately.
Industry Implications
1. Social Media - Social media platforms can make use of misogyny-detecting algorithms to filter out abusive content and create safer online environments for their users.
2. Online Content Moderation - Online content moderation companies can integrate AI-based text classification algorithms to enhance their detection capabilities and make their services more efficient.
3. Artificial Intelligence - The use of AI in developing new algorithms for contextual analysis and language detection can lead to disruptive innovation in fields like natural language processing and speech recognition technology.
1.2
Score
Popularity
Activity
Freshness
