ArguMiner
Our prototype extracts arguments from discussions to improve decision-making by providing the essence of the conversation.
Challenge
Mining Arguments
The challenge of argument mining involves extracting and analyzing arguments from textual sources, such as discussions or debates, to identify and understand the underlying claims, evidence, and reasoning. By uncovering and organizing arguments, argument mining aims to provide insights, facilitate information synthesis, and support decision-making processes.
About the prototype
We developed a custom code using an AI model from Hugging Face that specializes in argument recognition. Our process involves extracting text from uploaded files, segmenting it into sentences, and evaluating each sentence using the AI model to determine if it represents an argument. Similar arguments are then grouped together through clustering. We further employed another existing model to classify arguments as pro, con, or neutral. The result is a comprehensive table displaying indexed arguments and their corresponding scores. To visually represent our concept, we created a Figma prototype. The application allows users to upload files, insert article links, or conduct general topic searches. It presents an overview of arguments sorted by ratings, along with graphical representations of the argument-to-fluff ratio and rating distribution. By combining advanced AI models, efficient categorization algorithms, and an intuitive user interface, our prototype aims to empower individuals to quickly access and comprehend arguments, enabling them to make well-informed decisions and shape their own perspectives.
Outputs
Further information about the progress, milestones, and roadblocks.
Example for prototype
Example for prototype
Example for prototype
Impressions
